Serum 25(OH)D concentration & cardiovascular disease risk …rx... · 2019-02-13 · SERUM 25(OH)D CONCENTRATION & CARDIOVASCULAR DISEASE RISK ASSCOCIATIONS AMONG OLDER PUERTO RICANS

SERUM 25(OH)D CONCENTRATION & CARDIOVASCULAR DISEASE RISK

ASSCOCIATIONS AMONG OLDER PUERTO RICANS

Dissertation Presented

By

Aziza A. Jamal-Allial, M.S.

To

The Bouvé Graduate School of Health Sciences

in Partial Fulfillment of the Requirements for or Doctor of Philosophy in Pharmaceutical

Sciences with specialization in

Statistics & Epidemiology

NORTHEASTERN UNIVERSITY

BOSTON, MASSACHUSETTS

August, 2013

© 2013 Aziza A. Jamal-Allial

2

TABLE OF CONTENTS

ABSTRACT .................................................................................................................... 6

ACKNOWLEDGMENTS ............................................................................................... 9

DEDICATION .............................................................................................................. 11

LIST OF TABLES ........................................................................................................ 12

LIST OF FIGURES ....................................................................................................... 14

CHAPTER ONE ........................................................................................................... 15

INTRODUCTION ......................................................................................................... 15

STATEMENT OF THE PROBLEM ............................................................................ 17

REVIEW OF THE LITERATURE ............................................................................... 19

The Epidemiology of Cardiovascular Disease and Atherosclerosis ............................. 19

The Progression of Atherosclerosis ........................................................................... 20

VITAMIN D METABOLISM ...................................................................................... 23

Overview of Synthesis and Metabolism .................................................................... 23

Vitamin D Receptor: .................................................................................................. 24

The Association between Vitamin D Deficiency and CVD Risk Factors ................. 26

Proposed Mechanisms of Anti-Inflammatory Actions of Vitamin D ........................ 27

PRELIMINARY STUDIES .......................................................................................... 30

RESEARCH DESIGN AND METHOOLOGY ........................................................... 41

Study Population and Data Collection....................................................................... 41

Data Collection .......................................................................................................... 42

Covariates Assessments and Definition .................................................................... 43

Laboratory Assays ..................................................................................................... 44

Potential Covariates: .................................................................................................. 45

Data Entry and Quality Control ................................................................................. 46

STATISTICAL ANALYTICAL PLAN ....................................................................... 47

The Specific Aims ..................................................................................................... 49

BIBLIOGRAPHY ......................................................................................................... 51

CHAPTER TWO........................................................................................................... 60

3

Longitudinal Association of serum 25(OH)D concentration and Dyslipidemia among

Older Puerto Ricans: ..................................................................................................... 60

ABSTRACT .................................................................................................................. 61

INTRODUCTION ......................................................................................................... 63

MATERIALS AND METHODS .................................................................................. 65


Covariates Assessments and Definition .................................................................... 66

Laboratory assays ...................................................................................................... 67

Statistical analyses: .................................................................................................... 68

RESULTS...................................................................................................................... 70

DISCUSSION ............................................................................................................... 72

BIBLIOGRAPHY ......................................................................................................... 77

CHAPTER THREE ....................................................................................................... 92

An Inverse Association of Vitamin D Serum Concentration and Inflammatory Markers

....................................................................................................................................... 92

ABSTRACT .................................................................................................................. 93

INTRODUCTION ......................................................................................................... 95

MATERIALS AND METHODS .................................................................................. 96


Covariates and Exposure Assessments ...................................................................... 97

Laboratory assays ...................................................................................................... 99

Statistical analyses: .................................................................................................... 99

RESULTS.................................................................................................................... 101

DISCUSSION ............................................................................................................. 102

BIBLIOGRAPHY ....................................................................................................... 105

Poster Presentation: ................................................................................................. 116

PDF of the presentation of Chapter Two data at Experimental Biology 2013 ........... 116

CHAPTER FOUR ....................................................................................................... 117

Longitudinal Association of Vitamin D Serum Concentration & Adiposity Phenotypes

..................................................................................................................................... 117

ABSTRACT ................................................................................................................ 118

4

INTRODUCTION ....................................................................................................... 120

MATERIALS AND METHODS ................................................................................ 121

Study Population and Data Collection..................................................................... 121

Covariates Assessments and Definition .................................................................. 122

Laboratory assays .................................................................................................... 123

Statistical analyses: .................................................................................................. 124

RESULTS.................................................................................................................... 125

DISSCUSION ............................................................................................................. 127

BIBLIOGRAPHY ....................................................................................................... 130

Poster Presentation .................................................................................................. 138

CHAPTER FIVE ......................................................................................................... 139

The Association of Total and Regional Adiposity with Serum 25-hydroxyvitamin D

Concentration .............................................................................................................. 139

ABSTRACT ................................................................................................................ 140

INTRODUCTION ....................................................................................................... 142


Study Population and Data Collection..................................................................... 144

Covariates Assessments and Definition .................................................................. 145

Exposure Assessment: ............................................................................................. 146


Statistical analyses: .................................................................................................. 147

RESULTS.................................................................................................................... 148

DISCUSSION ............................................................................................................. 149

BIBLIOGRAPHY ....................................................................................................... 152

CHAPTER SIX ........................................................................................................... 164

DISCUSSION AND SUMMARY .............................................................................. 164

BIBLIOGRAPHY ....................................................................................................... 177

APPENDIX ................................................................................................................. 184

Serum Vitamin D Concentration, Dietary Sources, and its relation to BMI among

Puerto Ricans Living in the Greater Boston Area ....................................................... 185

5

ABSTRACT ................................................................................................................ 186

INTRODUCTION ....................................................................................................... 187


Study Population...................................................................................................... 188

Field data collection ................................................................................................. 189


Statistical analyses ................................................................................................... 191

RESULTS.................................................................................................................... 191

DISCUSSION ............................................................................................................. 194

BIBLIOGRAPHY ....................................................................................................... 197

6

ABSTRACT

Cardiovascular disease (CVD) is the number one single cause of mortality and

morbidity, accounting for approximately 30% of deaths worldwide. Numerous risk

factors for CVD are well established, but novel ones continue to emerge. Several reports

have suggested that there is an association between plasma 25-hydroxyvitamin D

concentration (25(OH)D) and CVD. Vitamin D deficiency is prevalent, especially among

ethnic minorities.

The objective of this dissertation was to investigate, cross-sectionally and

longitudinally, the relation between serum 25(OH)D concentration and CVD risk factors

in 970 Puerto Rican adults, 270 men and 700 women, aged 45-75 years: (1) to

investigate, longitudinally, the relation between serum 25(OH)D concentration and lipid

profile, specifically, high-density lipoprotein (HDL-C), low-density lipoprotein (LDL-C),

total cholesterol (TC) and triglycerides (TG). (2) to investigate, cross-sectionally, the

relation between serum 25(OH)D concentration and serum inflammatory markers,

specifically, tumor necrosis factor-alpha (TNF-α), interleukin-6 (IL-6), and high-

sensitivity C-reactive protein (hsCRP). (3) to investigate the longitudinal effect of body

mass index (BMI), and waist circumference (WC) as measures of adiposity on serum

25(OH)D concentration. (4) to investigate, cross-sectionally, the association of regional

and total adiposity, specifically, total body fat (TBF), percent body fat (%TBF), android

fat-mass (AFM) and gynoid fat-mass (GFM), using dual-energy x-ray absorptiometry

(DXA), on serum 25(OH)D concentration.

The cross sectional analyses from the two year time point showed that among all

participants, HDL-C had a significant positive association with plasma 25(OH)D

7

(P<0.01) while TG had significant negative association; but not LDL-C, or TC. The

longitudinal analyses showed that when predicting 2.5-year change in plasma HDL-C,

baseline-25(OH)D had a significant inverse association (β = -0.045 ± 0.02, P =0.007)

controlling for age, sex, baseline-HDL-C concentration, BMI, and follow-up time.

Similarly, after controlling for potential confounders, the baseline-25(OH)D was a

significantly positive predictor of the change in plasma TG concentration over time (β =

0.40 ± 0.02, P =0.03). With regard to inflammatory markers, after adjusting for age, sex,

and seasonality, the multivariate regression analyses showed that serum 25(OH)D had a

significantly negative association with TNF-α and IL-6, (P<0.01 and P<0.04),

respectively, but not with hsCRP. Further adjustment for numerous identified

confounders did not alter the significant associations between serum 25(OH)D

concentration with TNF-α and IL-6. The adiposity analyses showed that BMI and WC

were both negatively correlated, at both time points, with serum 25(OH)D concentration

(r = -0.1, P =0.02) and (r = -0.1, P =0.009), respectively. The multivariate analysis

showed that when predicting the change of serum 25(OH)D concentration, only baseline-

BMI had significantly inverse association (β = - 0.22, P =0.01) controlling for age, sex,

and baseline-BMI. This association remained significant after adjusting for numerous

recognized confounders, while baseline-WC was not significantly associated with serum

25(OH)D concentration. When investigating regional adiposity, our analyses showed that

neither TBF in grams nor %TBF were significantly associated with serum 25(OH)D.

However, AFM seems to have an inverse significant association with serum 25(OH)D,

whereas GFM showed a positive significant association with serum 25(OH)D.

8

In conclusion, even though 25(OH)D serum concentration was positively

associated with HDL-C in our cross-sectional analyses, the longitudinal analyses

showed that higher 25(OH)D serum concentration did not translate into an increase of the

atheroprotective HDL-C, nor a decrease in the not-atheroprotective plasma TG in this

population, which seems contradictory to the some of the literature . When we

investigated the association of serum 25(OH)D and inflammation, lower serum 25(OH)D

concentration was significantly associated with higher inflammatory concentration in this

population. The third major finding of the present dissertation work is that obesity has an

opposing effect on serum 25(OH)D concentration in this population of older Puerto Rican

adults living in the Boston area, which might accelerate vitamin D deficiency status.

Longitudinal studies are needed to clarify causal pathways.

9

ACKNOWLEDGMENTS

No words neither sentences can fully express my ample gratitude to everyone who was

there for me. I would like to extend my warmest appreciation and thankfulness to my committee

members, Dr. Mansoor Amiji, Dr. Katherine L. Tucker, Dr, John L. Griffith, Dr. S. John Gatley

and last but not least Dr. David R. Janero. This dissertation work would have not been achievable

without the endless support, encouragement, mentorship, and expertise I have endured from each

and every member of my committee.

My special appreciation is to Dr. Tucker for the tremendous and amazing opportunity she

gave me through these six years. Katy was there for me not only during the happy times, but as

well through the most difficult times with endless believe in my ability to achieve my dreams.

She understood that I am a mother first, and student second. Katy taught me how to be an

independent thinker and use my intellectual freedom as well as have a broad knowledge. I hope I

can carry a fraction of that knowledge I gained while working with her to the next chapter in my

career.

When I arrived at Northeastern I needed a home, and Dr. Amiji’s group was my home. I

cannot thank Dr. Amiji and his team enough for their generosity. His lab not only offered me

home, but also allowed me to broaden my knowledge and look at vitamin D from a different

perspective. My special thanks to Dr. Amit Singh, Dr. Srinivas Ganta, Dipti Deshpande, Shardool

Jain and Vanessa Nascimento among others who were there to mentor and answer my questions.

This study would not be possible without the tremendous mentorship of Dr, John L. Griffith. I

appreciate that every Tuesday at 1:00 pm, I know this is my blocked hour to ask, discuss and

converse with him about statistical modeling. I am better statistician because of his dedications. I

thank Dr. Gatley for all of these wisdom talks we had, it made me a better researcher. Finally, Dr.

Janero was always there to edit my writing, it made me a better writer.

10

This dissertation work would not be possible without the staff and the participants of the

Boston Puerto Rican Health Study. Without their dedication to the study, it would not be

possible. Much gratitude to the faculty and staff of the Pharmaceutical Science department and

Northeastern University, they have generously given their time and expertise to make us better

researchers. Every single interaction with them has been an amazing experience and I am truly

honored.

My deep gratitude to Dr. Mohamed Aleissa, the Cultural attaché at the Culture Mission of

Saudi Arabia, and Dr. Abudlaziz Aleisa, my advisor, among all the employee of SACM, for their

supports and encouragement. SCAM not only provided financial support but emotional

reassurance.

My friends were there for me all along, not only when laughing, but rather crying,

celebrating and even when it was hurtful, my special gratitude to Shilpa Bhuapathiraju, PhD.

Latrice Landry, PhD, Carlota Dao, PhD, Grace Bennett, PhD, Dr. Ons Khadra and Abeer Radady

M.D. In addition, special thanks to my amazing friend Stephanie Peterson for her editorial work

as well as to Shilpa Saklani for her amazing SAS macros.

Finally, I came to the most personal source of appreciation and gratefulness, my family.

My parents are the reason I have come to what I am today, they raised my siblings and I to

achieve and reach high in the sky. To my siblings, Mohammed, Olaa and May, who continued to

support and inspire me at every step of my life. Finally, to the most enduring source of strength

and motivation of my success, to my husband, Ashhal Jastania and my children, Hamza and

Mawada, this work is dedicated to you; you have given me endless love and support and without

you in my life, this was not possible. LOVE YOU!!

11

DEDICATION

To my lifelong partner, my husband Ashhal, and my rock and amazing children, Hamza

and Mawada, for the love, the devotion, the support, the persistence and the

encouragement needed to keep me going,

to my family, my friends, to everyone for truly believing in me.

This dissertation is dedicated to you!!

12

LIST OF TABLES

Table 1 List of studies of calcitriol and atherosclerosis using Apoe-/- mice....................29

Table 2 Descriptive characteristics of the BPRHS participants........................................35

Table 3 Selected dietary data of the participants of the BPRHS.....................................36

Table 4 Major Dietary Sources of Vitamin D intake among Puerto Rican adults…….37

Table 5 The Age-adjusted means for all variables……………………………………..38

Table 6 Multiple linear regression stratified by sex…………………………………...40

Table 7 Summary of calculated outcome measures and covariates……………………49

Table 1.1 Descriptive characteristic of men and women in the BPRHS & BPROS….....82

Table 1.2 Pearson partial linear correlation of serum 25(OH)D concentration and

adiposity phenotype at baseline and 2-year…………………………………………..…84

Table 1.3 Regression coefficients from longitudinal linear regression analysis of lipid

profile components and baseline serum 25(OH)D concentration………………………85

Table 1.4 Regression coefficients from longitudinal linear regression analysis of lipid

profile components and baseline serum 25(OH)D concentration………………………87

Table 2.1 Descriptive characteristic of the men and women of BPROS………………108


inflammatory markers………………………………………………………………….110

Table 2.3: Regression coefficients from multivariate regression analysis of inflammatory

markers and 25(OH)D concentration. …………………………………………………111

TABLE 2.4: Regression coefficients from multivariate regression analysis of hsCRP and

serum 25(OH)D concentration. ………………………………………………………..112

Table 3.1 Descriptive characteristic of men and women of BPRHS & BPROS…........133


adiposity phenotype at baseline and 2-year……………………………………………134

Table 3.3 Regression coefficients from longitudinal linear regression analysis of

adiposity phenotypes on plasma 25(OH)D concentration………………………..……135

Table 4.1 Descriptive characteristics of men and women of BPROS……………….....159


adiposity phenotype at 2-year……………………………………………..……………161

Table 4.3 Regression coefficients from cross-sectional regression analysis of TBF and

%TBF on plasma 25(OH)D concentration….…………………………….……………162

13

Table 4.4 Regression coefficients from cross-sectional regression analysis of AFM and

GFM on plasma 25(OH)D concentration……………………………….…………...…163

Table 6.1 Descriptive characteristics of men and women in the BPRHS………..…….211

Table 6.2 Age-adjusted means of serum 25(OH)D stratified by sex for different

characteristics among BPRHS participants……….………………..…….……….……212

Table 6.3 Correlates of serum 25(OH)D concentration among Puerto Rican adults

stratified by sex. ……………………………………………………………….……….214

Table 6.4 The least-square means of serum 25(OH)D concentration for the effect of

vitamin D supplements categories by sex………………………………………………215

Table 6.5 The least-square means of serum 25(OH)D concentration and seasons

categories by age groups…………………………………………………………..……216

Table 6.6 Major Dietary Sources of Vitamin D intake among Puerto Rican adults…..217

14

LIST OF FIGURES

Figure 1 Ergocalciferol structure D2………………………………………………….....23

Figure 2 Cholecalciferol structure D3…………………………………………………...23

Figure 3 Calcidiol Structure 25(OH)D………………………………………………….23

Figure 4 Calcitriol structure 1α-25(OH)2D……………………………………………..24

Figure 5 The percent of BPRHS women and men serum 25(OH)D concentration…….31

Figure 6 The association between vitamin D intakes from food & supplement………..32

Figure 7 The association between vitamin D intakes from food only sources…………32

Figure 8 Diagram illustrates the relationship between serum 25(OH)D and the various

CVD risk factors ………………………………………….……………………….…….50

Figure 1.1 Prevalence of low serum 25(OH)D concentration among PR………………89

Figure 1.2 The association between serum (25)OH at baseline and the change in HDL

over time modified by smoking………………………………………………………… 90

Figure 1.3 The association between serum (25)OH at baseline and the change in LDL

over time modified by smoking………………………………………………………… 91

Figure 2.1 Adjusted means of IL-6 concentration by tertiles of adjusted serum

25(OH)D………………………………………………………………………………..113

Figure 2.2 Adjusted means of TNF-α concentration by tertiles of adjusted serum

25(OH)D……………………………………………………………………………..…114

Figure 2.1 Adjusted means of hsCRP concentration by tertiles of adjusted serum

25(OH)D……………………………………………………………………………..…115

Figure 3.1 The monthly distribution of serum 25(OH)D for the participants of BPRHS at both time points……………………………………………………………………….136 Figure 3.2 The adjusted means of serum 25(OH)D within the categories of class III obesity…………………………………………………………………………….…….137 Figure 4.1 The illustrations of android region and gynoid region………………….….157 Figure 4.2 the association between serum 25(OH)D concentration with gynoid and

android fat among………………………………………………………………………158

Figure 5.1 Diagram illustrates the relationship between serum 25(OH)D and the various

CVD risk factors among the participants of the BPRHS…………………………….…177

Figure 6.1 Prevalence of low serum 25-hydroxyvitamin D concentration among Puerto

Rican men and women……………………………………………………………….....218

Figure 6.2 Association between the median intake of total vitamin D (defined by the

RDA) and serum 25(OH)D for BPRHS participants…………………………………...219

Figure 6.3 Association between vitamin D intake from food only sources and serum

25(OH)D among non-supplement users of the BPRHS participants…………………..220

Figure 6.4 Mean serum 25(OH)D concentration by BMI categories of Puerto Rican

women…………………………………………..………………………………………221

15

CHAPTER ONE

INTRODUCTION

Cardiovascular disease (CVD) is the number one single cause of mortality and

morbidity, accounting for approximately 30% of deaths worldwide in 2008. Numerous

risk factors for CVD are well established, but novel ones continue to emerge.

Atherosclerosis underlies various adverse vascular events that contribute to the

development of CVD. It is considered a chronic inflammatory disease of the muscular

arterial walls that is characterized by the accumulation of lipoproteins, immune response

activation, and increased expression of growth factors, all of which leads to endothelial

dysfunction and induces proliferation and migration of smooth muscle cells within the

arterial wall. Despite substantial advances in drug discovery and development for

treatment of cardiovascular disease over the last decade, the use of regimens such as

lipid-lowering agents, anti-platelet drugs, and anti-inflammatory drugs has not solved the

complex, multifactorial interplay of the pathophysiological drivers of CVD.

We proposed an interdisciplinary study of the association of various CVD risk

factors and serum 25-hydroxyvitamin D [25(OH)D] concentration of the participants of

Boston Puerto Rican Health Study (BPRHS) and the Boston Puerto Rican Osteoporosis

Study (BPROS). We examined the association between serum 25(OH)D concentration

and pro-inflammatory markers such interleukin-6 (IL-6), tumor necrosis factor-α (TNF-

α), and high-sensitivity C-reactive protein (hsCRP); plasma lipid profile, including total

16

cholesterol (TC), low-density lipoproteins (LDL-C), high-density lipoproteins (HDL-C),

and triglycerides (TG); and adiposity measures including: body mass index (BMI), waist

circumference (WC), and total body fat (TBF). The Boston Puerto Rican Health Project

is an ongoing longitudinal study with two time-points and a sample size of 1502

participants.

17

STATEMENT OF THE PROBLEM

Cardiovascular disease (CVD) is a major public health concern worldwide [1].

The list of traditional risk factors for CVD published by the American Heart Association

(AHA) includes family history, age, sex, behavioral/lifestyle, smoking, obesity, diabetes,

hyperlipidemia and hypertension [2]. Atherosclerosis is one of major causes of CVD and

it is characterized by lipid accumulation and systematic inflammation, and endothelial

dysfunction. The expression of inflammatory cytokines is a critical factor influencing

CVD progression [3]. Vitamin D (vitamin D) deficiency is now believed to be associated

with CVD and immune response: vitamin D appears to have a regulatory influence on the

inflammatory cytokines involved in the progression of systematic inflammation and

endothelial dysfunction [4].

The aims of our study: Puerto Rican men and women, aged 45-75 years, will be

evaluated both cross-sectionally and longitudinally, at baseline and at 2 years.

Aim One: To evaluate longitudinal associations between serum 25(OH)D concentration

and plasma lipid profile, precisely, high-density lipoproteins (HDL-C), low-density

lipoproteins (LDL-C), and total cholesterol (TC) and triglycerides (TG).

Aim Two: To evaluate cross-sectional associations between serum 25(OH)D

concentration and inflammatory markers, precisely, interleukin-6 (IL-6), tumor necrosis

factor-α (TNF-α), and high sensitivity C-reactive protein (hsCRP).

Aim Three: To evaluate the longitudinal associations between serum 25(OH)D

concentration and adiposity phenotype, specifically, body mass index (BMI) and waist

circumference (WC) among 970 participants of the BPRHS and BPROS studies

18

Aim Four: To evaluate cross-sectional associations between serum 25(OH)D

concentration and adiposity measures, specifically, total body fat mass (TBF) measured

by dual-energy x-ray absorptiometry (DEXA).

The relationship between serum vitamin D concentration and CVD risk factors in

general, and atherosclerosis, in particular, is well documented [5, 6]. The continued high

incidence of CVD-related mortality and morbidity emphasizes the need to distinguish not

only the contributors to CVD risk, but also to identify new therapeutic modalities for the

treatment of atherosclerosis as a primary cause of CVD. We believe that the proposed

work will contribute to the current knowledge of vitamin D and CVD risk. We know of

no studies that have reported on associations between serum vitamin D concentration and

CVD risk factors among older Puerto Ricans. Thus, results from our cohort project will

be the first to assess this association, longitudinally, within this particular ethnicity.

19

REVIEW OF THE LITERATURE

The Epidemiology of Cardiovascular Disease and Atherosclerosis

Despite significant progress in scientific research of CVD, it remains the leading

cause of morbidity and mortality worldwide [1, 7, 8]. According to the World Health

Organization (WHO), in 2010, CVD accounted for 17.3 million deaths, which is about

29% of all global deaths; it is projected that CVD will be responsible for ~24 million

deaths annually by the year 2030 [1]. In the USA, the mortality data for 2007 showed that

CVD remained the primary cause of death in the USA; approximately 34% of deaths

were CVD related [7, 9]. CVD risk factors are diverse, ranging from traditional factors

such as age, sex, smoking, inactivity, obesity, family history, hypertension, and

hyperlipidemia [10, 11]. However, traditional factors cannot fully explain the increase in

CVD incidence, and the emergence of more novel pathophysiological factors such as

inflammation, oxidative stress and adipocyte dysfunction might provide more insights on

the development of CVD [9]. Atherosclerosis is a pathological condition that underlies

various adverse vascular events that contribute to CVD development [12]. It is

characterized by accelerated accumulation macrophages and oxidized LDL (oxLDL)

within arterial blood vessels [3, 13]. The early stages of atherosclerosis are characterized

by endothelial dysfunction—impairment of endothelium-dependent vasodilatation, which

can be measured by brachial artery flow-mediated dilation (FMD) [14, 15].

20

The Progression of Atherosclerosis

The walls of large musular arteries consist of three distinct, well-differentiated

layers. The outermost layer relative to the vessel lumen is the tunica adventitia, which is

composed of connective tissue; inside this is the tunica media, which consists of smooth

muscle cells and elastic tissue; and the innermost layer, called the intima, consists of the

internal elastic lamina and a monolayer of endothelial cells, which is continuously and

directly exposed to blood and its constituents, along with the hemodynamic shear forces

associated with the flowing blood [3, 13]. The endothelial layer of the intima is a critical

controller of vascular homeostasis and functions as an anti-thrombotic vessel lining [16].

Vascular endothelial cells synthesize and release a wide array of mediators that act in

autocrine and/or paracrine modes to modulate the structure of the artery, smooth muscle

quiescence, endothelial passivity as well as blood constituents, vessel tone, and

hemodynamic reactivity.

The development of atherosclerosis involves early endothelial activation generally

believed to be triggered by various local factors such as oxidative stress and increased

expression of pro-inflammatory mediators, all of which lead to increases of the

adhesiveness of leukocytes or platelets through the increased expression of intracellular

adhesion molecule (ICAM). Another hallmark of endothelial activation is the increased

permeability of lipoproteins such as LDL-C and TG [3, 12, 13] and a defect in the

production of nitric oxide (NO) from L-arginine by endothelial NO synthase (eNOS)

[14].

21

NO is the most potent vasodilator yet identified and further acts to inhibit platelet

activity and smooth-muscle cell proliferation. In addition to mediating endothelium-

dependent vasodilatation, NO is critical for preventing platelet and leukocyte aggregation

as well as leukocyte adhesion and infiltration [12, 14]. Further, trapped LDL-C undergoes

modification including oxidation, lipolysis, and proteolysis, generating oxLDL-C, which

has the ability to inhibit the production of NO as well as to stimulate the endothelium to

produce pro-inflammatory molecules including adhesion molecules and growth factors.

Subsequently, monocytes and lymphocytes migrate across the endothelial layer into the

intima, where they proliferate and differentiate into macrophages. This migration is

triggered by the accumulation of oxLDL-C. Moreover, increased expression of ICAM

induces the adhesion of these monocytes along with LDL-C. Later, the extensive

modification of the oxLDL-C allows their recognition by the scavenger receptors within

the macrophage, and, hence, their uptake leads to foam-cell formation. Such receptor

expression is mediated by TNF-α and interferon-γ (IFN-γ) cytokines. The progressive

accumulation of foam cells, oxLDL-C, and monocytes leads to fatty streak formation.

As fatty streaks progress to advanced lesions, smooth muscle cells (SMC) migrate

to the intima along with T cells, subsequently leading to the development of a necrotic

core. This advanced lesion is characterized by dead foam cells, accelerated lipoprotein

and SMC macrophage accumulation, and enhanced inflammatory response, all of which

are covered by a fibrous cap that extends into the lumen. Rupture of the fibrous cap

exposes the content of the necrotic core to the blood, leading to coagulation and

recruitment of platelets, resulting in thrombosis [3, 12, 13].

22

Inflammation plays a pivotal role in all stages of atherogenesis, from foam cell to

plaque formation, and all the way to rupture and ultimately to thrombosis [17, 18]. In

fact, atherosclerotic lesions are characterized by the release of various pro-inflammatory

cytokines, such as hsCRP, IL-1β, IL-6 and TNF-α. The binding of LDL-C to the

endothelium is mediated by both IL-1β and TNF-α. Inflammatory cytokines such as IFN-

γ activate the macrophages to undergo programmed cell death, apoptosis, causing a mass

of extracellular lipids and debris. Thus, this cycle that starts with inflammation,

modification of lipoprotein systems and further inflammation is an illustration of how the

inflammatory response has a profound effect not only the development of atherosclerosis,

but in some cases it is considered a healing and repair response.

Emerging evidence suggests that serum vitamin D concentration are associated

with a wide spectrum of CVD risk factors, generally, and with atherosclerosis

development, specifically [19-25]. It has been hypothesized that higher serum vitamin

D concentration may play a major role in the attenuation of atherosclerosis

development and improvement of endothelial function. However, due to the complexity

of CVD, with its wide range of risk factors as well as its clinical biomarkers, and

complexity of vitamin D with its limited dietary sources vs. synthesis from sunlight, the

supporting evidence is mixed.

23

VITAMIN D METABOLISM

Overview of Synthesis and Metabolism

Vitamin D is a secosteriod compound with 27 carbons,

there are two forms of vitamin D, D2 or ergocalciferol (Figure 1),

and D3 cholecalciferol (Figure 2). It is an essential nutrient, but is

unlike most other nutrients in terms of its limited food sources

[26]. Only a few food sources, such as fatty fish and cod liver oil,

naturally contain vitamin D, making it difficult to achieve optimal

vitamin D concentration through food intake alone. Since the

1930s, dietary fortification with vitamin D has been established in

a number of different products, including milk, soymilk, orange

juice and selected cereal grains [27, 28]. In most cases, achieving

optimum vitamin D only with dietary intake is not feasible [26, 29-

31]; rather, a relatively efficient way of obtaining vitamin D is

endogenously, via the ultraviolet-β activation of 7-

dehydrocholestrol to pro-vitamin D3, within the epidermal basal

layer, which then undergoes isomerization producing the D3 form,

cholecalciferol (Figure 2). Afterward, cholecalciferol binds to the vitamin D-binding

protein (VDBP) and is transported to the liver, where it gets its first hydroxylation at the

25 position by 25-hydorxylase, producing calcidiol or 25(OH)D (Figure 3) [32].

Afterward, cholecalciferol binds to the vitamin D-binding protein (VDBP) and is

Figure 1: Ergocalciferol structure (D2)

Figure 2: Cholecalciferol

structure (D3)

Figure 3: Calcidiol

Structure (25(OH)D)

24

transported to the liver, where it gets its first hydroxylation at the 25 position by 25-

hydorxylase, producing calcidiol or 25(OH)D (Figure 3) [32].

Several cytochrome-P450 enzymes (CYP-450s), such as

CYP2J2, and CYP3A4, also exhibit hydroxylase activity, which

metabolizes the less potent D2 form (Figure 1) [33]. 25(OH)D is

the blood biomarker for determining the status of vitamin D: it is

characterized by a relatively long half-life of about 3 weeks.

Mainly in the kidney, a second hydroxylation occurs at the 1-α

position by CYP27B1, a 1α-hydorxylase, converting 25(OH)D to 1α,25(OH)2D, calcitriol

(Figure 4), which is the active metabolite, but which has a very short half-life, up to only

12-15 hours [32, 34-36]. This pathway is exquisitely regulated by parathyroid hormone

(PTH), serum calcium concentration(Ca+), and phosphate (Pi) concentration, and also by

negative feedback inhibition from 1α,25(OH)2D concentration [37]. CYP-P450 regulates

the catabolism of vitamin D through the actions of CYP24A1, 24-hydroxylase, which

hydroxylates both 25(OH)D and 1α,25(OH)2D at the 24 position, leading to the inactive

hydrophilic metabolites 24,25(OH)2D and 1α,24,25(OH)3D, which are readily excreted

in the urine; consequently, this catabolic pathway is a limiting step [34].

Vitamin D Receptor:

Vitamin D and its analogs exert their differential effects on target cells through

two main pathways: genomic and non-genomic pathways. The genomic pathway is

mediated through the binding of 1α,25(OH)2D and/or its analogs to the nuclear vitamin D

Figure 4: Calcitriol

structure (1α-25(OH)2D)

25

receptor (VDR) to exert transcriptional activation and gene expression repression. VDR

is classified as a nuclear hormone receptor and is expressed and localized in various

tissues where vitamin D has regulatory the pancreatic β cells, monocytes, dendritic cells,

and peripheral T cells. The binding of 1α,25(OH)2D to VDR requires a specific ligand,

retinoid-X receptor (RXR). Thus, this later compound binds to response elements

(VDREs) in the promoter region of targeted genes, which is followed by a cascade of

binding of transcription factors, co-activators, and co-repressors to increase the

transcription and/or suppress of target genes. The non-genomic pathway is through the

binding of 1α,25(OH)2D and/or its analogs to cytosolic and membrane VDR, which

initiates a phosphorylation cascade of mitogen-activated protein kinases (MAPK) and

extracellular signal-regulated kinase (ERK), which might participate in a cross-talk with

the genomic pathway to modulate gene expression [37-39]. In conclusion, VDR is

considered a ligand-induced receptor that controls the expression of more than a thousand

genes, many of whom are involved in cell division, differentiation, and growth.

The literature is rich with considerable data for the role for vitamin D in bone and

mineral homeostasis [40, 41], however; in the late 1970’s, the discovery of VDR among

various types of tissues, other than bone and small intestine, has provided new insights to

the non-traditional roles of vitamin D in the immune system regulation [42], cancer [43-

45], CVD [46, 47] and diabetes [48, 49]. Moreover, the cloning of CYP27B1 in various

tissues has provided more evidence that vitamin D activation is not only limited to the

functions including: intestine, bone tissues and kidneys, as well as some immune cells,

namely, kidney, but rather 1α-hydroxylase is locally expressed where vitamin D is

needed including: (1) regulating the proliferation of normal cells such as the

26

keratinocytes; (2) regulating hormone secretion such as PTH, insulin, and growth

hormone; (3) regulating both adaptive and innate immunity [50].

The Association between Vitamin D Deficiency and CVD Risk Factors

Vitamin D deficiency is considered a global epidemic. It has been estimated that

more than one billion people have vitamin D concentration that are considered,

minimally, insufficient [26, 29, 51-53]. Thus far, there is no agreement within the medical

and/or scientific area on defining optimal levels of 25(OH)D. However, most scientists

define sufficient serum 25(OH)D concentration as above 75 nmol/L; insufficient as 50-75

nmol/L; deficient as below 50 nmol/L; and severely deficient as below 25 nmol/L [26, 54-

56]. In the US, 48% of young white girls in Maine and 52% of Hispanic adolescents in

Boston were found to be vitamin D deficient [57, 58]. Another report indicated that

among Hispanics, Puerto Ricans had the highest prevalence of vitamin D deficiency, 26%

[59]. Interestingly, another report investigated the levels of vitamin D concentration

among Puerto Ricans in Puerto Rico and found that about 55% of their sample had

vitamin D concentration >70 nmol/L, 31% had levels between 50 and 70 nmol/L, and

14% had levels below 50 nmol/L [60]. In the Middle East, vitamin D deficiency ranges

between 30 to 50% among adults [26]. A recent report from Scotland indicated that high

percentages of healthy adults were vitamin D deficient, 29%, or severely deficient, 34.5%

[61].

Aside from increased risk of osteoporosis and lower bone mineral density [40],

vitamin D deficiency has been associated with increased risk of several illness including

CVD and hypertension [20, 23, 62]; cancer, including breast and prostate [45, 63];

27

obesity and related metabolic disorders such as type-2-diabetes [48, 64]; sarcopenia and

muscle weakness [65, 66]; and multiple sclerosis [67]. In the last decade, a wide range of

scientific evidence has linked vitamin D deficiency to increased risk of CVD. First, a

number of large, nonrandomized, prospective studies have documented that vitamin D

deficiency was associated with CVD morbidity and mortality. The Framingham

Offspring Study showed that the incidence of CVD was higher among individuals whose

serum 25(OH)D concentration was below 37 nmol/L [68]. The Luric study showed that

participants whose 25(OH)D concentration was in the lowest quartile (median=19

nmol/L) or the second lowest (median=33 nmol/L) had high adjusted risk for CVD

mortality, relative to those in the highest quartile (median=71 nmol/L), 122% and 82%

risk, respectively [69]. A meta-analysis showed that vitamin D supplementation, ranging

between 300 and 2000 IU/day, was associated with lower total mortality, relative risk

(RR) = 0.93 (95% CI:0.87-0.99) [70].

Proposed Mechanisms of Anti-Inflammatory Actions of Vitamin D

Several studies have documented that vitamin D exerts anti-proliferative and pro-

differentiating effects on many types of normal and malignant cells [71, 72]. One of the

proposed mechanisms for the association of calcitriol with atherosclerosis is through the

increased expression of MAP kinase phosphatase-5 (MKP5) which, in turn,

dephosphorylates and, thereby, inactivates MAPKs, such as p38 and the stress-activated

protein kinase Jun-N-terminal kinase (JNK). This action leads to a decrease in the

production of pro-inflammatory cytokines such as IL-6 [4]. Another proposed mechanism

28

is through inhibition of nuclear-factor-κB (NF-κB) activation and signaling, which is a

major modulator of inflammatory and immune response, through increasing the

expression of IκB protein and activating IKB kinase (IKKB) [73]. The activation of

NFκB is accompanied with increased expression of a wide range of inflammatory

cytokines, such as IL-6, IL-8, and TNFα [73, 74]. One recent study, by Talmor and

collaborators, showed that in vitro calcitriol lowered the activity of NF-κB and p38, and

the expression of ICAM and IL-6 [75]. Table 1 illustrates several in vivo experimental

murine studies using ApoE-/- mice that investigated the effects of different vitamin D

analogs on atherosclerosis.

The association between vitamin D and adiposity, in general, and body fat, in

particular, was explained by Lumb and colleagues in 1971. They were the first to

introduce the hypothesis of vitamin D sequestration in adipose tissue and muscle [76].

Later, they were able to substantiate that prior hypothesis and confirm the link between

adiposity and vitamin D [77]. Subsequently, Wortsman and colleague confirmed the

aforementioned premise, stating that due to vitamin D deposition within adipose tissue,

the bioavailability of serum 25(OH)D and 1,25(OH)2D is lowered, leading to vitamin D

insufficiency/deficiency and higher PTH. They showed that obese people evidence, lower

[48].

29

Table 1: List of studies of calcitriol and analogs on atherosclerosis using ApoE-/- mice

Author Hypothesis Treatme

nt Animals

Treatmen

t Dose

Becker et, al. 2011 [78]

aortic wall *19-nor *1,25D3

ApoE-/- (n=16/trmt)

IP5x /wk for 10 wks

*19-norc�0.1 ug/kg

*1,25D3 �0.03 µg/kg

Husain et, al. 2010 [79]

atherosclerosis *9-nor

*Enalapril

ApoE-/- females

n=12/trmt

IP 3x/wk for 16 wks

*19-norc�200 ng

*Enlp�30mg/kg

Takeda et, al. 2010 [80]

atherosclerosis 1,25D3

only ApoE-/- n=6-9/trmt

Oral 2x/wk for

12 wks

20 ng 200 ng (10 ug/kg)

Wang et, al, 2011

Oxidative stress and proinflamm

cytokines 1αD2

NON mice (males)

n=12/trmt

IP for 28 wks

1αD2 �125 ng/kg

Wang et, al, 2011 [81]

Oxidative stress and proinflamm

cytokines 1αD2

NON mice males

n=12/trmt

IP for 28 wks

1αD2 �125 ng/kg

30

PRELIMINARY STUDIES

The BPRHS is an ongoing longitudinal study that follows 1505 older Puerto

Ricans in the Greater Boston Area. For the analyses we are presenting here, we used the

baseline and 2-year data time points. Complete baseline serum 25(OH)D concentration

and dietary vitamin D intake assessments were available for 1432 participants, 1004

women (mean age =57.3 years) and 428 men (mean age = 56.6 years). Figure 13 shows

the prevalence of serum 25(OH)D concentration among the participants of BPRHS. The

majority of these Puerto Rican men and women (~75%) were considered vitamin D

deficient, with concentration below 50 nmol/L, whereas only 4.2-3% had sufficient serum

25(OH)D concentration of vitamin D, with levels above 75 nmol/L.

Women had significantly lower scores for serum albumin, serum creatinine,

systolic and diastolic blood pressure, HCY, VLDL-C, TG, and physical activity, but

significantly higher scores for BMI, HDL-C, TC, LDL-C, and poverty rate. Compared to

men, more women are taking supplements with a higher content of vitamin D -above 200

international units (IU) as well higher dietary intake of energy-adjusted total vitamin D

relative to men. No differences were observed between men and women for glycosylated

hemoglobin, fasting glucose, waist circumferences or Homa-IR (Table 2).

Analysis of dietary data shows that Puerto Rican women have significantly higher

carbohydrate intake, with more of their calories deriving from carbohydrates compared to

Puerto Rican men. On the other hand, the women tend to have significantly lower total

energy and dietary cholesterol intakes than do Puerto Rican men; also, a lower portion of

the women’s energy derives from fat sources compared to that for Puerto Rican men.

31

However, no significant differences were found by sex when dietary fiber and total fat

were compared (Table. 3). Figures 6 and 7 show the baseline association between serum

concentration and deciles of median intake of vitamin D from food and supplements

among all participants (n=1416). There was a wide range of vitamin D intake, 201 to 599

IU/day, when vitamin D supplements are included with food sources (Figure 6). On the

other hand, the range was very narrow when considering only the food sources of vitamin

D, 208 to 217 IU/day (Figure 7). Food products made from whole milk, 2% milk, and 1%

milk constituted 34% of vitamin D sources due to fortification with vitamin D. Fish

ranked second, constituting 16.2% of the vitamin D source. Eggs and ready-to-eat cereal

were fourth and fifth respectively, contributing about 4.5% and 4.3% of dietary vitamin D

intake (Table 3).

Figure 5: The percent of BPRHS women and men in two age groups, with serum 25(OH)D concentration (n=1495): deficient (<50 nmol/L), insufficient (50-75 nmol/L) or sufficient (>75 nmol/L).

0

10

20

30

40

50

60

70

80

40-60 yrs 60-75 yrs 40-60 yrs 60-75 yrs

Women Men

Per

cen

tage

Deficient

Insufficient

Sufficient

32

Table 4 shows the breakdown of age-adjusted means of serum 25(OH)D

concentration by different population factors. In general, the use of vitamin D-containing

supplements was associated with higher overall serum 25(OH)D concentration: our

results show that both Puerto Rican men and women who consume minimal daily

supplements (200 IU) vitamin D had significantly greater serum 25(OH)D concentration

compared to non-users: 50.1 and 47.0 nmol/L, respectively (P for trend <0.0001). We

observed a significant decreasing trend concentration across the four seasons, starting

from the summer toward spring season (P for trend <0.0001). Puerto Rican women aged

60-75 years had significantly higher serum 25(OH)D concentration compared to those

aged 45-59 yr: 45 and 42.1 nmol/L, respectively (P-value 0.009). On the other hand,

women with higher BMI had significantly lower serum 25(OH)D concentration (P for

trend 0.05). In addition, past and current female smokers had significantly lower

25(OH)D concentration (P for trend 0.04) compared to those who have never smoked.

30

35

40

45

50

55

20

8.5

20

9.5

21

0

21

0.5

21

1

21

1.5

21

2

21

3

21

4

21

7

Se

rum

25

(OH

)D n

mo

l/L

Median Vitamin D intake IU/day by deciles

from food sources only

Figure 6 & 7: The association between vitamin D intake from food & supplement (left) and food only sources (right) and serum 25(OH)D concentration from baseline (n=1416) BPRHS participants. The LS-means adjusted for age, sex, and seasonality are plotted against the median for each decile group of energy-adjusted vitamin D intake.

30

35

40

45

50

55

20

9

21

0

21

1

21

1.5

21

3

21

5

21

9.5

39

9.5

40

3

59

9.5

Se

rum

25

(OH

)D n

mo

l/L

Median Vitamin D intake IU/day by

deciles from food & supplement sources

33

Moreover, women, but not men, showed a significant increasing trend in serum 25(OH)D

concentration with increased physical activity score (P for trend 0.02). Among men,

lower concentration of serum albumin was associated with significantly lower serum

25(OH)D concentration: 32 and 44 nmol/L, respectively (P-value 0.05). Although higher

alcohol intake appeared to correlate with lower serum 25(OH)D among men as well as

women, this did not reach statistical significance. Finally, there was no significant

relationship between serum 25 (OH) D concentration, and activities of daily living (ADL)

for either men or women.

Known and potential factors that influence serum 25(OH)D concentration were

used to construct linear regression models (Table 5). The multiplicative interaction of sex

with age was highly significant (P<0.0001), thus models are presented separately for men

and women. In the fully adjusted model for women, serum 25(OH)D concentration was

significantly lower, by 0.15 nmol/L, for each year of age (P = 0.03). Lower concentration

were observed in winter and spring, -5.0 and -8.4 nmol/L, respectively (P= 0.0006 and <

0.0001) relative to summer. Higher BMI was associated with significantly lower

concentration, - 0.21 nmol/L (P< 0.0001 for each BMI unit). Serum 25(OH)D

concentration was 0.8 nmol/L greater for each IU difference in dietary vitamin D

(P<0.0001). Use of vitamin D-containing supplements was independently associated with

higher serum concentration (P<0.0001). Supplements containing vitamin D in the range

of 200-400 IU/day or > 400 IU/day, were associated with significant higher serum

25(OH)D concentration—to 5.0 and 11.1 nmol/L respectively—vs. no supplement use

(P<0.0001). Serum albumin was also associated with higher serum vitamin D

concentration (P= 0.008). Women who never smoked had significantly higher serum

34

concentration, 3.4 nmol/L compared to currently smoking women (P =0.01). Serum

creatinine, physical activity score, and alcohol intake were not significantly associated

with serum 25(OH)D concentration in women..

For men, seasonality was associated with similar decreasing serum 25(OH)D

concentration as seasons moved away from the summer season, but the reduction was

more pronounced than in women: the winter and spring levels were -11.6 and -12.3

nmol/L lower, respectively, (P< 0.0001 for each difference). Supplements with 200-400

IU/day of vitamin D (but not those >400 IU/day) were associated with serum 25(OH)D

concentration 9.0 nmol/L higher than non-users concentration (P<0.0001). In addition, as

reported above for women, serum albumin was independently associated with higher

serum 25(OH)D concentration (P= 0.003). In men, age, BMI, dietary intake of vitamin

D, and smoking did not contribute to serum 25(OH)D concentration, unlike changes seen

in women; nor was there any associations with the serum creatinine, physical activity

score, or alcohol intake.

35

TABLE 2 Descriptive characteristics of the BPRHS participants a

Covariates Women

(n=1048)

Men

(n=447) P-valueb

Age (years) 57.3 ± 7.5 56.6 ± 8.0 0.08b

Education (below 8th grade %) 48.0 44.0 0.1 c

Poverty (%) 62.0 51.0 0.0002 c

BMI (kg/m2) 33.0 ± 7.0 29.6 ± 5.2 <0.0001b

Waist (cm) 102.0 101.5 0.7

Serum Albumin c (g/L) 42.5 ± 7.0 43.2 ± 3.6 <0.0001b

Serum Creatinine c (nmol/L) 68.0 ± 27.5 87.5 ± 45.1 <0.0001b

Number of Walking Miles/day b 1.0 ± 1.4 1.3 ± 1.6 <0.0001b

Serum 25-(OH)Dc (nmol/L) 43.1 ± 16.0 43.5 ± 16.5 0.6b

Vitamin D Supplement use (> 200 IU %) 30.0 24.0 0.02c

Dietary vitamin D e.f (IU/day) 212 ± 3.0 212 ± 3.4 0.5b

Total vitamin D e,f (IU/day) 285 ± 4.5 285 ± 4.3 0.01b

Glycosylated Hemoglobin (%) 0.07 ± 2 0.07 ± 2 0.8

Fasting Glucose (mg/dL) 120± 53.0 121 ± 52.5 0.7

Homa-IR 6.03 ± 17 6.0 ± 10.2 0.004 b

LDL-C (mg/dL) 110.4 ± 34 101.0 ± 35.2 <.0001b

HDL-C (mg/dL) 47.02 ± 12.1 40.5 ± 12.4 <.0001b

VLDL-C (mg/dL) 29.3 ± 15.0 32.0 ± 19 0.007b

TG (mg/dL) 157.0 ± 99 174.0 ± 140 0.007b

Total Cholesterol (mg/dL) 188.2 ± 41 174.2 ± 43 <.0001b

Homocysteine (µmol/L) 8.45 ± 3.7 10.37 ± 5.7 <.0001b

Systolic blood pressure (mm Hg) 134.1 ± 19.2 137.1 ± 19 <.0001b

Diastolic blood pressure (mm Hg) 80.11 ± 10.4 83.0 ± 11.2 <.0001b

a Values are presented as mean ± SD unless indicated b

P-value are based on T-test, c P-value are based on chi-square test

d Range = 38.3 Total vitamin D= vitamin D from supplement + vitamin D from diet. e The recommended dietary allowance (RDA) for vitamin D for both men and women aged 19-50 years, and 51-70 years = 600 IU (15 ug /day) [82] f Adjusted for total energy

36

TABLE 3: Selected dietary intake data for the participants of the BPRHS

aMean ± standard error adjusted for age and total energy. bSignificant differences by sex (t-tests)

Women (n=1048)

Men (n=447)

P-valueb

Total Energy (kcal) 2021± 32 2367 ± 52 < 0.001b Carbohydrate (g) 273 ± 2 264 ± 3 < 0.01b

Dietary fiber (g) 18.5 ± 0.2 18.3 ± 0.3 0.06

Total Fat (g) 75.9 ± 0.5 77.0 ± 0.8 0.09

Saturated fat (g) 22.9 ± 0.2 23.1 ± 0.4 0.07

Monounsaturated fat (g) 26.2 ± 0.2 26.6 ± 0.3 0.5

Polyunsaturated fat (g) 17.7 ± 0.2 18.1 ± 0.3 0.07

Protein (g) 91.8 ± 0.7 91.2 ± 1.2 0.7

Cholesterol (mg) 298± 5 334 ± 7 < 0.001b

% Energy from Carbohydrate 52.2 ± 0.3 50.1 ± 0.5 < 0.001b

% Energy from Total Fat 32.0 ± 0.2 32.8 ± 0.3 < 0.05b

% Energy from Protein 17.4 ± 0.1 17.2 ± 0.2 0.07

37

TABLE 4: Major Dietary Sources of Vitamin D intake among Puerto Rican adultsa

Rank Food Groups % of dietary vitamin D

intake

1 Milk (whole, 2% &1% milk products)

32.4

2 Supplements 28.7 3 Fish 16.2 4 Eggs 4.5 5 Cereal, Ready-to-eat, Cold 4.3 6 Beef 2.3

7 Processed Meat, Sausage, Frank 2.1 8 Chicken/Turkey 2.0 9 Pork 1.3 10 Ice cream, Sherbet, Frozen Yogurt 1.2 11 Yogurt 1.1

12 Cheese 0.9 a Only food groups that contributed more than 0.8% of dietary vitamin D intake

38

TABLE 5 Age-adjusted means for all variables

Variables Women Men

n (%)* LSM SEM P value n (%) LSM SEM P

value

Age (y)

45-59 651 (61.5%)

42.0 0.64 281 (63.0%)

44.4 1.00

60-75 407 (38.5%) 45.0 0.81 0.007 164 (37.0%)

42.0 1.33 0.1

Vitamin D Supplement IU/day) † †‡

≤ 200 738 (70.0%) 41.1 0.6 338 (77.3%)

41.5 0.8

200 - 400 224 (21.5%) 46.5 1.1 <.0001 90 (20.5%)

51.0 1.6 <0.0001

≥400 77 (8.5%) 53.1 2.0 <.0001 9 (2.1 %) 39.2 5.36 0.9

<.0001|| <0.0001

||

Season of measurement†

Summer (July-Sept)

356 (34.5%) 44.5 0.9 137 (32.0%)

46.5 1.4

Fall (October- Dec)

264 (26.0%) 46.0 0.8 0.4 124 (29.0%)

48.7 1.4 0.6

Winter (Jan-March)

184 (18.0%) 41.1 1.2 0.05 78 (18.0%)

37.5 1.8 <0.0001

Spring (April-June)

224 (21.5%) 38.2 1.1 <.0001 91 (21.3%)

37.0 1.6 <0.0001

<.0001|| <0.0001

||

Body Mass Index (kg/m2) †

≤ 25 114 (11.0%) 44.0 1.5 80 (18.5%)

43.5 2.0

25-30 282 (27.3%) 45.5 0.9 0.8 164 (37.5%)

44.0 1.3 0.9

30-35 307 (29.3%) 44.5 0.9 0.9 130 (30.2%)

43.0 1.5 0.9

≥ 35 307 (29.3%) 44.5 0.9 0.08 130 (30.2%)

43.0 1.5 0.9

0.001 || 0.9

||

Smoking Status†

39

No 832 (79.5%) 44.0 1.1 291 (66.0%)

44.0 1.0

Yes 217 (20.5%) 41.0 0.5 0.01 149 (34.0%)

43.0 1.4 0.6

Alcohol Intake†

No 675 (64.5%) 43.0 0.7 211 (49.0%)

44.5 1.2

Yes 376 (35.5%) 44.0 0.8 0.4 228 (51.0%)

43.0 1.1 0.2

Serum Albumin† (g/L)

< 38 49 (5%) 42.0 2.2 16 (4.0%)

35.0 4.0

≥ 38 964 (95%) 43.1 0.5 0.5 410 (96.0%)

44.0 0.8 0.03

Serum Creatinine (nmol/L) †•

35.5 - 61.0 280 (27.5%) 41.5 1.0 19 (4.5%)

34.5 4.0

62.0 -70.0 298 (29.0%) 42.5 1.0 0.6 46 (11.0%)

40.0 2.4 0.3

71.0 -79.5 223 (22.0%) 45.3 1.1 0.009 92 (21.5%)

42.0 2.0 0.1

80.0 - 698 221 (22.0%) 44.1 1.1 0.1 270 (63.0%)

45.3 1.0 0.01

0.02 || 0.009

||

Miles Walked/day†

0 mile/day 263 (25.0%) 41.3 1.1 52 (12.0%)

40.5 2.4

0- 0.5 mile/day 395 (38.0%) 42.5 0.8 0.5 140 (32.0%)

44.0 1.4 0.3

> 0.5 mile/day 367 (37.0%) 45.0 1.0 0.01 247 (56.3%)

44.0 1.1 0.3

0.01 || 0.4

||

1P-value for individual comparisons to the referent group that is based on Dunnett

multiple testing. 2 P-value for Trend.*Quintiles; **Quartiles

3 Reference group, P-values are calculated with Dunnett adjustment for multiple comparisons.

40

TABLE 6 Multiple linear regression, stratified by sex (n= 1542)

Variable Women (n=966) Men (n=404)

β SE P-value β SE P-value

Age 0.15 0.07 0.03 -0.13 0.1 0.2

Seasons (Fall vs. summer) -1.6 1.3 0.1 -3.0 2.0 0.1

(Winter vs. Summer) -5.0 1.4 0.0006 -11.6 2.3 <0.0001

(Spring vs. Summer) -8.40 1.3 <0.0001 -12.3 2.2 <0.0001

BMI -0.21 0.07 0.006 0.02 0.15 0.8

vitamin D supplement use

(200-400 IU vs. None)

5.0 1.1 <0.0001 9.0 2.0 <0.0001

(> 400 IU vs. None) 11.1 2.0 <0.0001 -1.0 5.2 0.8

Energy-adjusted dietary vitamin D

(IU/day)

0.8 0.2 <0.0001 0.25 0.2 0.2

Serum Creatinine (nmol/L) 0.01 0.01 0.5 0.01 0.01 0.4

Serum Albumin (g/L) 0.4 0.1 0.008 0.64 0.2 0.003

Physical activity score 0.15 0.12 0.2 -0.10 0.13 0.4

Smoking (Never vs. Current) 3.4 1.3 0.01 -0.5 2.0 0.7

(Past vs. Current) 2.2 1.5 0.1 0.15 2.0 0.9

Alcohol (Never vs. Heavy) 3.8 2.5 0.1 3.6 2.3 0.1

(Moderate vs. Heavy) 4.6 2.5 0.07 3.2 2.4 0.1

41

RESEARCH DESIGN AND METHOOLOGY

The work presented in this dissertation involves collaboration among

investigators with expertise in nutritional epidemiology and biochemistry, and statistical

methodologies for clinical research. Using epidemiologic methods, we aimed to

understand the association of between vitamin D status and CVD risk factors among

Puerto Rican adults in Boston.

The parent study for the epidemiologic analyses is the Boston Puerto Rican

Health Study, an ongoing longitudinal study to examine the associations among social

and psychological stress, nutrition, allostatic load, and measures of depression, cognitive

function, and physical disability in a cohort of older Puerto Ricans, aged 45-75 years,

living in the Greater Boston, MA. Consenting participants returned to the Human

Nutrition Research Center on Aging (HNRCA) at Tufts University at the end of 2 years

for measurements of bone and body composition. For this analysis there were 1505

observations from baseline, 1276 observations from a 2-year time point, and 970 bone

and body composition measurements.

Study Population and Data Collection

Participants were identified from areas of high Hispanic density in the Boston

metropolitan area, as specified in the year 2000 Census. Blocks with 10 or more Hispanic

persons aged 45 and older were selected for enumeration from census tracks that contain

at least 25 Puerto Rican adults in the specified age range. After complete block

enumeration (blocks were visited up to 5 or 6 times, and on weekends and evenings to

42

maximize identification of all relevant households), households with at least one Puerto

Rican adult between the ages of 45-75 years were identified; one qualified individual was

randomly selected from each household, and invited to participate. Only those

participants who were unable to answer questions due to serious health conditions or who

were not able to provide reliable information for other reasons such as dementia or

alcoholism, were excluded from participation [83].

Data Collection

Baseline and 2-year visit: Informed consent was obtained during the baseline

interview, and neuropsychological tests were conducted to identify participants who

might require assistance (e.g., from a proxy); participants were excluded if their mini

mental state examination score was <10. Enrolled participants completed additional

questionnaires to capture information on general background and socioeconomic status,

dietary intake, food security, health history, health behaviors, health insurance, medical

conditions and use of medication, activities of daily living instrument, migration status,

stress, depressive symptomology, and cognitive function. Moreover, anthropometric and

blood pressure measures, and physical function tests, were completed. Participants were

provided with urine collection containers, two saliva tubes with procedures for collection,

and instructions for the next day’s 12-hour-fasting blood draw. The next day, a

phlebotomist collected 12-hour fasting blood, urine, and saliva samples. Samples were

analyzed at the Nutrition Evaluation Laboratory (NEL) at Tufts University; blood

samples were immediately cooled to 4oC and plasma separated within 4 hours in a

refrigerated centrifuge.

43

Bone visit protocol: During the 2-year follow-up visit, an appointment was made

for participants to visit the HNRCA to obtain bone density and body composition

measurements and an additional blood draw, and to complete additional questionnaires

on sun exposure, family history of osteoporosis, and history of fracture. Participants were

provided with written instructions for the urine collection, taking PABA tablets, and for

fasting. Two days before their scheduled visit, participants received a reminder phone call

and reinforced instructions for urine collection and overnight fasting. They were asked to

complete the 24-hour urine collection in two containers, one for each 12 hours. All

laboratory measurements for this study were performed by the NEL.

Covariates Assessments and Definition

Serum 25(OH)D concentration was measured with an 125I radioimmunoassay kit

(DiaSorin Inc, Stillwater, MN 55082) as specified by the manufacturer (68100E). Overall

adiposity was assessed from whole-body DXA scans (DXA; Lunar Prodigy; GE

Healthcare, Madison, WI) measured during the bone visit. Regional body composition

measures (android fat mass (grams and %), gynoid fat mass (grams and %), and

android/gynoid fat mass ratio were assessed with use of ENCORE software, version 12.2.

Weight, and waist and hip circumferences were measured in duplicate. Weight was

measured using a clinical scale (Toledo Weight Plate, Model I5S, Bay State, and Systems

Inc. Burlington, MA). Knee height was measured with a Harpenden pocket stadiometer

(also used for standing height). Knee height was used to estimate standing height for

participants who could not stand, or who had a spinal curvature. Body mass index (BMI)

was calculated as weight (kg) divided by height (m) squared [84, 85].

44

Laboratory Assays

Serum high sensitivity C-reactive protein (hsCRP) was measured with a solid-

phase, enzyme-labeled chemiluminescentimmunometric assay (Immulite 1000,

Diagnostic Products Corporation (DCP) Los Angeles, CA 90045-5597), as specified by

the manufacturer (PILKCR-7, 2003-11-25). The intra- and inter-assay CV% were 6.0%

and 7.3% respectively.

Serum TNF-α & IL-6 were each measured by non-cross-reacting enzyme-linked

immunoassays (ELISA), with use of monoclonal and polyclonal antibodies for the

analysis of specific cytokine antigens (Quantikine ELISA, R&D Systems, Minneapolis,

MN, USA). At the NEL, for IL-6, the intra-assay CV% is (1.6-4.2%) and the inter-assay

CV% is (3.3-6.4%). For TNF-α, the intra-assay CV% is 5.3-8.8% and the inter-assay

CV% is (10.8-16.7%). Plasma cholesterol was measured with EDTA plasma using an

enzymatic endpoint reaction in a Cobas Mira Analyzer with Roche Diagnostics Item

#44307. Ciba Corning Normal and Abnormal assayed control serum (Items #9702 and

#9705) and Sigma Diagnostics elevated lipid control (Item #L1008) were used as controls

[86]. HDL-C (mg/dL) was analyzed using EDTA plasma with the enzymatic endpoint

reaction in a Cobas Mira Analyzer with Roche Diagnostics (Item #44120), and was

calibrated with a Roche Diagnostics Cobas Calibrator with Roche Diagnostic Cholesterol

Reagent (Item#44307) Dextran Sulfate-Magnesium Precipitating Reagent. The standard

was Roche Diagnostics Cobas Calibrator Serum, (Item #44157), and the controls were

Sigma Assayed Bi-Level HDL-C Cholesterol Controls, (Items #H4009 and #H5009)

(Allan, Poon et al. 1974; Warnick, Benderson et al. 1982). LDL-C was calculated with

use of the Friedewald formula, except in the cases when triglyceride (TG) concentration

45

exceed 400 mg/dL [87]. TG was analyzed with an enzymatic endpoint reaction on the

Olympus AU400e with Olympus Triglyceride Reagents (OSR6033) (Olympus America

Inc., Melville, NY).

Potential Covariates:

Potential covariates include age (continuous), sex (categorical), socio-

demographic parameters such as education (categorical), income (continuous), smoking

status (categorical), alcohol intake (categorical), total energy intake (continuous),

physical activity (continuous /categorical). Information was obtained from all participants

on the following variables: age, education, and household income. This questionnaire was

designed based on questions from NHANES III [88], HHANES [89], the National Health

Interview Survey Supplement on Aging [90], and refined during our previous field

experience. Health-related behaviors: Information about current and past use of alcohol

and smoking and about physical activity were obtained. Seasonality was defined as

follows: July to September was coded as summer, October to December as fall, January

to March as winter, and April to June as spring.

Diabetes was considered present if fasting blood glucose exceeded 125 mg/dL or

if the participant was being treated with insulin or oral hypoglycemic agents. Serum

glucose was measured at baseline and 2 years, using an enzymatic, kinetic reaction on the

Olympus AU400e with Olympus Glucose Reagents (Olympus America Inc., Melville,

NY) at the NEL.

46

Data Entry and Quality Control

Participants were coded with a subject identification number. All data were

entered and managed at the BPRHS center at Northeastern University on password-

protected computers. Data entry reports were generated periodically to detect and

minimize missing data and data entry errors. When missing data were identified in

questionnaires, participants were contacted by phone within two weeks to retrieve the

information. Data were backed up each day and stored on our central server using unique

subject identification numbers.

Power and sample size:

The BPRHS study is ongoing and currently has baseline and 2 year-data through

home visits on 1502 and 1276, participants, respectively. In addition, we have bone and

body composition measurements on 970 subjects. The proposed sample size estimates

were selected to assure that we have at least 90% power to detect expected associations

that are reliable but consistent with published studies. We derived our sample size

estimates and the estimated power from the standard normal approximation to the Fisher

logarithmic transformation for linear correlations, which are likely conservative given

that multiple linear regression models generally provide greater precision, smaller

standard errors, and greater power.

47

STATISTICAL ANALYTICAL PLAN

Analyses were conducted with SAS© version 9.3 (SAS Institute Inc., Cary, NC).

Formal hypothesis testing was two-sided, with a nominal type I error of 0.05. We

assessed the modeling assumptions of non-linearity, and potential outliers and influential

data were considered. Plots and graphs were used to summarize the distributions of

primary analytic variables and other descriptive data. For all aims, results were

summarized using regression estimates, p-values, and 95% CI. Geometric means are

reported for log-transformed variables. Analyses were adjusted for covariates, including

sex, age, seasonality, etc. Potential confounding and effect modifiers were checked for by

including interaction term into the regression model. For continuous data, independent

sample t-tests were employed; for categorical data, chi-square test for differences in

proportions was employed. Only participants who met the eligibility criteria for the

intervention and with complete data for the primary analytical variables were included in

this analysis. Table 7 lists variables included in our statistical models.

48

TABLE 7 Summary of calculated outcome measures and covariates

Variables Measurement

Factor

Technique

Predictor variable

Exposure Serum 25(OH)D concentration (nmol/L)

Continuous /categorical

ELISA

Outcome variables

Pro-inflammatory cytokines

IL-6 (mg/L) TNF-α hsCRP

continuous 10 20 20

ELISA

Adiposity BMI WC TBF

continuous 10 20 20

Measurement tools DXA

Lipid profile

HDL-C (mg/dL) TC LDL-C TG

continuous

10 20 20 20

Lipid concentration

Covariates

Other blood measure

Albumin (g/L) Creatinine (µmol/L) Parathyroid hormone White blood cell counts Seasonality (blood drawn)

Continuous/ categorical

Kinetics

Demographics

Age Sex education poverty

categorical/ continuous

Designated questionnaire & measurement tools

Health and Life Style Factors

Smoking (y/n) Alcohol (gm/day) physical activity level # Blocks walking/day

categorical/ continuous

CVD

Systolic BP Diastolic BP Heart disease CVD medications

continuous categorical

Vitamin D intake Vitamin D dietary Vitamin D supplement

continuous FFQ

49

The Specific Aims

Specific Aim One: To evaluate the association between baseline serum 25(OH)D

concentration (nmol/L) and lipid profile constituent concentration (mg/dL), namely,

HDL-C, LDL-C, TG, and TC over time.

Primary Hypothesis One: lower serum 25(OH)D concentration is associated with 2.5

year change in HDL-C concentration.

Specific Aim Two: To evaluate the association between serum 25(OH)D concentration

with IL-6, TNF-α, and hsCRP.

Primary Hypothesis Two: lower serum 25(OH)D concentration is associated with higher

circulating levels of inflammatory markers.

Specific Aim Three: to evaluate the association between serum 25(OH)D concentration

and adiposity.

Primary Hypothesis Three: lower serum 25(OH)D concentration is associated with

higher adiposity as measured by BMI, and as a secondary aim, by waist circumference

(WC).

Specific Aim Four: to evaluate the cross-sectional association between serum 25(OH)D

concentration with total and regional adiposity

Primary Hypothesis Four: lower serum 25(OH)D concentration is associated with

higher total and regional adiposity measures as measured by dual-energy x-ray

absorptiometry (DXA), including total body fat (TBF).

50

FIGURE 8 Diagram illustrates the relationship between serum 25(OH)D and the various

CVD risk factors

51

BIBLIOGRAPHY

1. Organization, W.H., Cardiovascular diseases (CVDs), in Fact sheet N°3172011,

WHO: Geneva.

2. Balagopal, P., et al., Nontraditional Risk Factors and Biomarkers for

Cardiovascular Disease: Mechanistic, Research, and Clinical Considerations for

Youth. Circulation, 2011. 123(23): p. 2749-2769.

3. Lusis, A.J., Atherosclerosis. Nature, 2000. 407(6801): p. 233-241.

4. Krishnan, A.V. and D. Feldman, Mechanisms of the anti-cancer and anti-

inflammatory actions of vitamin D. Annual review of pharmacology and

toxicology, 2011. 51: p. 311-36.

5. Tomaschitz, A., et al., Independent association between 1,25-dihydroxyvitamin D,

25-hydroxyvitamin D and the renin–angiotensin system: The Ludwigshafen Risk

and Cardiovascular Health (LURIC) study. Clinica Chimica Acta, 2010. 411(17–

18): p. 1354-1360.

6. Vacek, J.L., et al., Vitamin D Deficiency and Supplementation and Relation to

Cardiovascular Health. The American Journal of Cardiology, 2012. 109(3): p.

359-363.

7. Sherry L. Murphy, B.S.J.X., M.D.; and Kenneth D. Kochanek, M.A., Deaths:

Preliminary Data for 2010, in National Vital Statisictis Report2012, CDC:

Division of Vital Statistics.

8. Writing Group, M., et al., Heart disease and stroke statistics--2010 update: a

report from the American Heart Association. Circulation, 2010. 121(7): p. e46-

e215.

9. Members, W.G., et al., Executive Summary: Heart Disease and Stroke Statistics—

2011 Update. Circulation, 2011. 123(4): p. 459-463.

10. Wilson, P.W. and B.F. Culleton, Epidemiology of cardiovascular disease in the

United States. American Journal of Kidney Diseases, 1998. 32(5 Suppl 3).

11. Lloyd-Jones, D.M., et al., Lifetime risk for developing congestive heart failure:

the Framingham Heart Study. Circulation, 2002. 106(24): p. 3068-72.

52

12. Tedgui, A. and Z. Mallat, Cytokines in Atherosclerosis: Pathogenic and

Regulatory Pathways. Physiological Reviews, 2006. 86(2): p. 515-581.

13. Ross, R., Atherosclerosis is an inflammatory disease. Am Heart J, 1999. 138: p.

S419-20.

14. Davignon, J. and P. Ganz, Role of Endothelial Dysfunction in Atherosclerosis.

Circulation, 2004. 109(23 suppl 1): p. III-27-III-32.

15. Lind, L., Lipids and endothelium-dependent vasodilation—A review. Lipids,

2002. 37(1): p. 1-15.

16. Almirall, J., et al., Association of low serum 25-hydroxyvitamin D levels and high

arterial blood pressure in the elderly. Nephrology Dialysis Transplantation, 2010.

25(2): p. 503-9.

17. Ameri, P., et al., High prevalence of vitamin D deficiency and its association with

left ventricular dilation: an echocardiography study in elderly patients with

chronic heart failure. Nutrition Metabolism & Cardiovascular Diseases, 2010.

20(9): p. 633-40.

18. Baz-Hecht, M. and A.B. Goldfine, The impact of vitamin D deficiency on diabetes

and cardiovascular risk. Current Opinion in Endocrinology, Diabetes & Obesity,

2010. 17(2): p. 113-9.

19. Bostick, R.M., et al., Relation of calcium, vitamin D, and dairy food intake to

ischemic heart disease mortality among postmenopausal women. American

Journal of Epidemiology, 1999. 149(2): p. 151-61.

20. Giovannucci, E., et al., 25-hydroxyvitamin D and risk of myocardial infarction in

men: a prospective study. Archives of Internal Medicine, 2008. 168(11): p. 1174-

80.

21. Oh, J., et al., 1,25(OH)2 Vitamin D Inhibits Foam Cell Formation and Suppresses

Macrophage Cholesterol Uptake in Patients With Type 2 Diabetes Mellitus.

Circulation, 2009. 120(8): p. 687-698.

22. Sugden, J.A., et al., Vitamin D improves endothelial function in patients with Type

2 diabetes mellitus and low vitamin D levels. Diabet Med, 2008. 25(3): p. 320-5.

53

23. Deshpande, D.D., D.R. Janero, and M.M. Amiji, Therapeutic strategies for

endothelial dysfunction. Expert opinion on biological therapy, 2011. 11(12): p.

1637-54.

24. Ridker, P.M., Clinical application of C-reactive protein for cardiovascular

disease detection and prevention. Circulation, 2003. 107(3): p. 363-9.

25. Shishehbor, M.H. and D.L. Bhatt, Inflammation and atherosclerosis. Current

Atherosclerosis Reports, 2004. 6(2): p. 131-9.

26. Holick, M.F., Vitamin D deficiency. N Engl J Med, 2007. 357(3): p. 266-81.

27. Armas, L.A., B.W. Hollis, and R.P. Heaney, Vitamin D2 is much less effective

than vitamin D3 in humans. Journal of Clinical Endocrinology & Metabolism,

2004. 89(11): p. 5387-91.

28. Holick, M.F., et al., The vitamin D content of fortified milk and infant formula. N

Engl J Med, 1992. 326(18): p. 1178-81.

29. Holick, M.F., High prevalence of vitamin D inadequacy and implications for

health. Mayo Clin Proc, 2006. 81(3): p. 353-73.

30. Holick, M.F., Resurrection of vitamin D deficiency and rickets. J Clin Invest,

2006. 116(8): p. 2062-72.

31. Holick, M.F., Deficiency of sunlight and vitamin D. Bmj, 2008. 336(7657): p.

1318-9.

32. Ramos-Lopez, E., et al., CYP2R1 (vitamin D 25-hydroxylase) gene is associated

with susceptibility to type 1 diabetes and vitamin D levels in Germans.

Diabetes/Metabolism Research and Reviews, 2007. 23(8): p. 631-636.

33. Rhieu, S.Y., et al., A new insight into the role of rat cytochrome P450 24A1 in

metabolism of selective analogs of 1alpha,25-dihydroxyvitamin D. Archives of

biochemistry and biophysics, 2011. 509(1): p. 33-43.

34. Prosser, D.E. and G. Jones, Enzymes involved in the activation and inactivation of

vitamin D. Trends in biochemical sciences, 2004. 29(12): p. 664-73.

35. Cheng, J.B., et al., Genetic evidence that the human CYP2R1 enzyme is a key

vitamin D 25-hydroxylase. Proceedings of the National Academy of Sciences of

the United States of America, 2004. 101(20): p. 7711-5.

54

36. Jones, G., Pharmacokinetics of vitamin D toxicity. The American Journal of

Clinical Nutrition, 2008. 88(2): p. 582S-586S.

37. Deeb, K.K., D.L. Trump, and C.S. Johnson, Vitamin D signalling pathways in

cancer: potential for anticancer therapeutics. Nature reviews. Cancer, 2007. 7(9):

p. 684-700.

38. Plum, L.A. and H.F. DeLuca, Vitamin D, disease and therapeutic opportunities.

Nat Rev Drug Discov, 2010. 9(12): p. 941-55.

39. POLLY, P., et al., VDR-Alien: a novel, DNA-selective vitamin D3 receptor-

corepressor partnership. The FASEB Journal, 2000. 14(10): p. 1455-1463.

40. Holick, M.F., Evolution and function of vitamin D. Recent Results Cancer Res,

2003. 164: p. 3-28.

41. Holick, M.F., The influence of vitamin D on bone health across the life cycle. J

Nutr, 2005. 135(11): p. 2726S-7S.

42. Etten, E.v. and C. Mathieu, Immunoregulation by 1,25-dihydroxyvitamin D3:

Basic concepts. The Journal of Steroid Biochemistry and Molecular Biology,

2005. 97(1–2): p. 93-101.

43. Bertone-Johnson, E.R., et al., Plasma 25-hydroxyvitamin D and 1,25-

dihydroxyvitamin D and risk of breast cancer. Cancer Epidemiol Biomarkers

Prev, 2005. 14(8): p. 1991-7.

44. Chen, T.C., et al., Prostatic 25-hydroxyvitamin D-1alpha-hydroxylase and its

implication in prostate cancer. J Cell Biochem, 2003. 88(2): p. 315-22.

45. Chen, T.C. and M.F. Holick, Vitamin D and prostate cancer prevention and

treatment. Trends Endocrinol Metab, 2003. 14(9): p. 423-30.

46. Judd, S.E. and V. Tangpricha, Vitamin D deficiency and risk for cardiovascular

disease. American Journal of the Medical Sciences, 2009. 338(1): p. 40-4.

47. Kim, D.H., et al., Prevalence of hypovitaminosis D in cardiovascular diseases

(from the National Health and Nutrition Examination Survey 2001 to 2004).

American Journal of Cardiology, 2008. 102(11): p. 1540-4.

48. Wortsman, J., et al., Decreased bioavailability of vitamin D in obesity. The

American Journal of Clinical Nutrition, 2000. 72(3): p. 690-693.

55

49. Liel, Y., et al., Low circulating vitamin D in obesity. Calcified tissue international,

1988. 43(4): p. 199-201.

50. Michael F, H.P.D., M.D., ed. Vitamin D. Physiology, Molecular Biology, and

clinical applications. Second ed. NUTRITION AND HEALTH, ed. P. Adrianne

Bendich2010, Springer Humana Press. 1155.

51. Bischoff-Ferrari, H.A., et al., Estimation of optimal serum concentrations of 25-

hydroxyvitamin D for multiple health outcomes. American Journal of Clinical

Nutrition, 2006. 84(1): p. 18-28.

52. Chapuy, M.C., et al., Prevalence of vitamin D insufficiency in an adult normal

population. Osteoporos Int, 1997. 7(5): p. 439-43.

53. Malabanan, A., I.E. Veronikis, and M.F. Holick, Redefining vitamin D

insufficiency. Lancet, 1998. 351(9105): p. 805-6.

54. Dawson-Hughes, B., et al., Estimates of optimal vitamin D status. Osteoporos Int,

2005. 16(7): p. 713-6.

55. Holick, M.F., Vitamin D status: measurement, interpretation, and clinical

application. Ann Epidemiol, 2009. 19(2): p. 73-8.

56. Holick, M.F. and T.C. Chen, Vitamin D deficiency: a worldwide problem with

health consequences. Am J Clin Nutr, 2008. 87(4): p. 1080S-6S.

57. Gordon, C.M., et al., Prevalence of vitamin D deficiency among healthy

adolescents. Arch Pediatr Adolesc Med, 2004. 158(6): p. 531-7.

58. Sullivan, S.S., et al., Adolescent girls in Maine are at risk for vitamin D

insufficiency. J Am Diet Assoc, 2005. 105(6): p. 971-4.

59. Araujo, A., et al., Serum 25-hydroxyvitamin D and bone mineral density among

Hispanic men. Osteoporosis International, 2009. 20(2): p. 245-255.

60. Palacios, C., et al., Determinants of vitamin D status among overweight and obese

Puerto Rican adults. Ann Nutr Metab, 2012. 60(1): p. 35-43.

61. Zgaga, L., et al., Diet, environmental factors, and lifestyle underlie the high

prevalence of vitamin D deficiency in healthy adults in Scotland, and

supplementation reduces the proportion that are severely deficient. Journal of

Nutrition, 2011. 141(8): p. 1535-42.

56

62. Michos, E.D., et al., 25-Hydroxyvitamin D deficiency is associated with fatal

stroke among whites but not blacks: The NHANES-III linked mortality files.

Nutrition, (0).

63. Freedman, D.M., et al., Serum Levels of Vitamin D Metabolites and Breast

Cancer Risk in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening

Trial. Cancer Epidemiology Biomarkers & Prevention, 2008. 17(4): p. 889-894.

64. Botella-Carretero, J.I., et al., Vitamin D deficiency is associated with the

metabolic syndrome in morbid obesity. Clinical Nutrition, 2007. 26(5): p. 573-80.

65. Oh, J.H., et al., The level of vitamin D in the serum correlates with fatty

degeneration of the muscles of the rotator cuff. Journal of Bone & Joint Surgery -

British Volume, 2009. 91(12): p. 1587-93.

66. Visser, M., et al., Low vitamin D and high parathyroid hormone levels as

determinants of loss of muscle strength and muscle mass (sarcopenia): the

Longitudinal Aging Study Amsterdam. Journal of Clinical Endocrinology &

Metabolism, 2003. 88(12): p. 5766-72.

67. Kragt, J., et al., Higher levels of 25-hydroxyvitamin D are associated with a lower

incidence of multiple sclerosis only in women. Multiple Sclerosis, 2009. 15(1): p.

9-15.

68. Wang, T.J., et al., Vitamin D deficiency and risk of cardiovascular disease.

Circulation, 2008. 117(4): p. 503-11.

69. Pilz, S., et al., Association of Vitamin D Deficiency with Heart Failure and

Sudden Cardiac Death in a Large Cross-Sectional Study of Patients Referred for

Coronary Angiography. Journal of Clinical Endocrinology & Metabolism, 2008.

93(10): p. 3927-3935.

70. Autier, P. and S. Gandini, Vitamin D supplementation and total mortality: a meta-

analysis of randomized controlled trials. Archives of Internal Medicine, 2007.

167(16): p. 1730-7.

71. Chung, I., et al., Anti-proliferative effects of calcitriol on endothelial cells derived

from two different microenvironments. The Journal of Steroid Biochemistry and

Molecular Biology, 2007. 103(3–5): p. 768-770.

57

72. Jacques M, L., Immunomodulatory actions of 1,25-Dihydroxyvitamin D3. The

Journal of Steroid Biochemistry and Molecular Biology, 1995. 53(1–6): p. 599-

602.

73. Berk, B.C., et al., Endothelial Atheroprotective and Anti-inflammatory

Mechanisms. Annals of the New York Academy of Sciences, 2001. 947(1): p. 93-

111.

74. de Vries-van der Weij, J., et al., Anti-inflammatory salicylate beneficially

modulates pre-existing atherosclerosis through quenching of NF-κB activity and

lowering of cholesterol. Atherosclerosis, 2010. 213(1): p. 241-246.

75. Talmor, Y., et al., Calcitriol blunts pro-atherosclerotic parameters through

NFkappaB and p38 in vitro. Eur J Clin Invest, 2008. 38(8): p. 548-54.

76. Becker, L.E., et al., Effect of paricalcitol and calcitriol on aortic wall remodeling

in uninephrectomized ApoE knockout mice. American Journal of Physiology -

Renal Physiology, 2011. 300(3): p. F772-F782.

77. Husain, K., et al., Effects of Paricalcitol and Enalapril on Atherosclerotic Injury

in Mouse Aortas. American Journal of Nephrology, 2010. 32(4): p. 296-304.

78. Takeda, M., et al., Oral Administration of an Active Form of Vitamin D3

(Calcitriol) Decreases Atherosclerosis in Mice by Inducing Regulatory T Cells

and Immature Dendritic Cells With Tolerogenic Functions. Arteriosclerosis,

Thrombosis, and Vascular Biology, 2010. 30(12): p. 2495-2503.

79. Bodyak, N., et al., Activated vitamin D attenuates left ventricular abnormalities

induced by dietary sodium in Dahl salt-sensitive animals. Proceedings of the

National Academy of Sciences of the United States of America, 2007. 104(43): p.

16810-5.

80. Wang, X.X., et al., Vitamin D receptor agonist doxercalciferol modulates dietary

fat-induced renal disease and renal lipid metabolism. American Journal of

Physiology - Renal Physiology, 2011. 300(3): p. F801-F810.

81. Lumb, G.A., E.B. Mawer, and S.W. Stanbury, The apparent vitamin D resistance

of chronic renal failure: A study of the physiology of vitamin D in man. The

American Journal of Medicine, 1971. 50(4): p. 421-441.

58

82. Mawer, E.B., et al., The distribution and storage of vitamin D and its metabolites

in human tissues. Clin Sci, 1972. 43(3): p. 413-31.

83. Dietary Reference Intakes for Calcium and Vitamin D. Committee to Review

Dietary Reference Intakes for Vitamin D and Calcium. , in 2010, Institute of

Medicine: Washington, DC: National Academy of Sciences.

84. Tucker, K., et al., The Boston Puerto Rican Health Study, a longitudinal cohort

study on health disparities in Puerto Rican adults: challenges and opportunities.

BMC Public Health, 2010. 10(1): p. 107.

85. Bermudez, O.I., E.K. Becker, and K.L. Tucker, Development of sex-specific

equations for estimating stature of frail elderly Hispanics living in the

northeastern United States. American Journal of Clinical Nutrition, 1999. 69(5):

p. 992-8.

86. Chumlea, W., et al., Stature prediction equations for elderly non-Hispanic white,

non-Hispanic black, and Mexican-American persons developed from NHANES III

data. J Am Diet Assoc, 1998. 98: p. 137 - 142.

87. Daughaday, W.H. and P. Rotwein, Insulin-like growth factors I and II. Peptide,

messenger ribonucleic acid and gene structures, serum, and tissue

concentrations. Endocr Rev, 1989. 10(1): p. 68-91.

88. Allain, C.C., et al., Enzymatic determination of total serum cholesterol. Clin

Chem, 1974. 20(4): p. 470-5.

89. Friedewald, W.T., R.I. Levy, and D.S. Fredrickson, Estimation of the

concentration of low-density lipoprotein cholesterol in plasma, without use of the

preparative ultracentrifuge. Clin Chem, 1972. 18(6): p. 499-502.

90. Dreon, D., et al., Use of NHANES Data to Assign Nutrient Densities to Food

Groups in a Multiethnic Diet History Questionnaire. Nutrition and Cancer, 1993.

20: p. 223 - 230.

91. Delgado, J.L., et al., Hispanic Health and Nutrition Examination Survey:

methodological considerations. Am J Public Health, 1990. 80 Suppl: p. 6-10.

59

92. Block, G. and A.F. Subar, Estimates of nutrient intake from a food frequency

questionnaire: the 1987 National Health Interview Survey. J Am Diet Assoc,

1992. 92(8): p. 969-77.

93. Tucker, K., et al., Adaptation of a food frequency questionnaire to assess diets of

Puerto Rican and non-Hispanic adults. American Journal of Epidemiology, 1998.

148(5): p. 507 - 518.

94. Brecher, G., M. Schneiderman, and G.Z. Williams, Evaluation of electronic red

blood cell counter. Am J Clin Pathol, 1956. 26(12): p. 1439-49.

95. Thomas, G.N., et al., Vitamin D Levels Predict All-Cause and Cardiovascular

Disease Mortality in Subjects With the Metabolic Syndrome. Diabetes Care, 2012.

35(5): p. 1158-1164.

96. McGill, A.T., et al., Relationships of low serum vitamin D3 with anthropometry

and markers of the metabolic syndrome and diabetes in overweight and obesity.

Nutr J, 2008. 7: p. 4.

60

CHAPTER TWO

Longitudinal Association of serum 25(OH)D concentration and Dyslipidemia

among Older Puerto Ricans:

The Boston Puerto Rican Health Study & Boston Puerto Rican Osteoporosis Study

Aziza Jamal-Allial1, John L. Griffith2,3, Katherine L. Tucker 3

1Department of Pharmaceutical Science. 3Department of Health Sciences, 2Bouvé

College of Health Sciences. Northeastern University.

Boston, MA, USA 02115

Presented at Experimental Biology 2013 (Oral Presentation)

61

ABSTRACT

Vitamin D deficiency is highly prevalent. Several reports suggest an association

between plasma 25-hydroxyvitamin D concentration (25(OH)D) and cardiovascular

disease (CVD). We examined, cross-sectionally and longitudinally, the relation between

serum 25(OH)D concentration and lipid profile of 970 Puerto Rican adults, 270 men and

700 women, aged 45-75 years. Cross sectional analyses at the two year time point

showed that high-density lipoprotein cholesterol (HDL-C) was significantly associated

with plasma 25(OH)D (P<0.01); but not low-density lipoprotein cholesterol (LDL-C),

triglycerides (TG), or total cholesterol (TC).

In longitudinal analyses, baseline-25(OH)D was significantly associated with

subsequent 2.5-year change in plasma HDL-C (β = -0.045 ± 0.02, P =0.007) controlling

for age, sex, baseline-HDL-C concentration, BMI, follow-up time, physical activity, total

energy, alcohol, smoking, seasonality, poverty, and education. This relationship remained

significant after adjustment for CVD medications and heart disease status. After

controlling for potential confounders, baseline-25(OH)D was a significantly positive

predictor of the change in plasma TG concentration over time (β = 0.40 ± 0.02, P =0.03).

In conclusion, even though 25(OH)D serum concentration was positively

associated with HDL-C in our cross-sectional analyses, the longitudinal analyses showed

that higher 25(OH)D serum concentration did not translate into an increase of the

atheroprotective HDL-C, nor a decrease in the not-atheroprotective plasma TG in this

population of older Puerto Rican adults living in the Boston area.

62

Abbreviations:

25-hydroxyvitamin D (25(OH)D); cardiovascular disease (CVD); confidence interval

(CI); United States (USA); food-frequency questionnaire (FFQ); high-density lipoprotein

(HDL-C); low-density lipoprotein (LDL-C); triglycerides (TG); total cholesterol (TC);

vitamin D receptor (VDR); cytochrome-P450 enzymes (CYP-450s).

63

INTRODUCTION

Vitamin D receptor (VDR) is found not only in not only in bone metabolism

tissues, but in a variety of tissues, making it likely that vitamin D actions extend beyond

maintaining calcium homeostasis and skeletal integrity [93, 94]. Achieving optimum

vitamin D status with only dietary sources is not generally feasible [26, 29-31]; rather,

endogenous synthesis is the most efficient way. This occurs via the ultraviolet-β

activation of 7-dehydrocholestrol producing the D3 form, cholecalciferol [32]. This later

binds to the vitamin D-binding protein (VDBP) and is transported to the liver, where it

gets its first hydroxylation at the 25 position by 25-hydorxylase, producing 25-

hydroxyvitamin D (25(OH)D). The second hydroxylation occurs at the 1-α position by

the actions of one of the cytochrome-P450 enzymes (CYP-450s), namely, CYP27B1, a

1α-hydorxylase, converting 25(OH)D to 1α,25(OH)2D, calcitriol. 25(OH)D is the blood

biomarker that determines the status of vitamin D; it is characterized by a relatively long

half-life of about 3 weeks. 1α,25(OH)2D is the active metabolite that binds to VDR, yet it

has very short half-life, up to only 12-15 hours [32, 34-36]. Vitamin D pathways are

exquisitely regulated by parathyroid hormone (iPTH), serum calcium (Ca+) and

phosphate (Pi) concentration, as well as by the negative feedback inhibition by

1α,25(OH)2D concentration [37, 95]. iPTH under the influence of low serum calcium

increases the second hydroxylation, thus, increasing 1α,25(OH)2D [95].

Notwithstanding increased risk of osteoporosis and lower bone mineral density

[40], vitamin D deficiency has been associated with increased risk of mortality [96] as

well as several chronic illnesses, including: CVD, hypertension [20, 23, 62], cancer [45,

63], obesity, type-2-diabetes [48, 64] and inflammatory disorders [67, 97, 98]. In the last

64

decade, a wide range of scientific evidence has linked vitamin D deficiency to an

increased risk of cardiovascular disease (CVD) [6, 25, 99, 100]. Dyslipidemia, which is

characterized by either low plasma HDL-C, high LDL-C and/or high TG concentration, is

a major risk factor for CVD concentration [101, 102]. Several epidemiological and

ecological studies have shown that lower 25(OH)D concentration is associated with an

unfavorable lipid profile [103-107]. However, divergent results were observed when

intervention studies investigated the effect of vitamin D supplementation on a number of

CVD risk factors that included the lipid profile [108].

Ethnic minorities, largely, and Puerto Ricans, explicitly, are reported to have a

greater burden of health disparities compared to non-Hispanic whites, thus increasing the

risk of developing chronic illness [109-111]. In their systematic review, Renzaho and

colleagures showed that ethnic minorities are substantially predisposed to a greater risk of

vitamin D deficiency among all age groups. They further conclude that a number of

indications link vitamin D deficiency to obesity-related chronic diseases [112].

The aim of this study is to determine the longitudinal association between serum

25(OH)D concentration and lipid profile components, namely, HDL-C, LDL-C, TC, and

TG (mg/dL), among participants of the Boston Puerto Rican Health Study (BPRHS) and

the Boston Puerto Rican Osteoporosis Study (BPROS), living in the Greater Boston Area.

65

MATERIALS AND METHODS


Description of the recruitment and methods of the BPRHS [83] and its ancillary

study, the BPROS [113], are described in detail elsewhere. Concisely, at both time

points, a bilingual interviewer conducted home visits and administrated detailed

questionnaires on general background, socioeconomic status, education, household

income, health and health behaviors that include medical diagnosis, detailed use of

medication(s), depression and a psychological acculturation scale. They also

administered questionnaires on dietary intake, food security, activities of daily living,

migration status, stress, depressive symptomology, and cognitive function. Moreover, the

interviewer collected information about anthropometrics, physical performance measures

and blood pressure. The following day, the study phlebotomist collected 24-hr urine, 12-

hr fasting blood and saliva. By December 2010, 1504 participants completed baseline

visits. Once the 2-yr follow up was completed, participants were re-consented for the

BPROS. By May 2013, 1257 participants completed 2year follow-up visits and 970 of

these completed the BPROS study. All study protocols were approved by the Institutional

Review Boards at Tufts Medical Center and Northeastern University, and each

participant provided written informed consent.

There were no significant differences in socio-demographic characteristics

between those with complete or incomplete information. The present study includes

complete longitudinal data for 700 Puerto Rican women and 270 Puerto Rican men.

66


At each time point, in addition to the designed questionnaires, physical function

tests were completed and anthropometric measures were taken. Blood pressure measures

were obtained at three times during the interview. Weight, and waist and hip

circumferences were measured in duplicate. Weight was measured using a clinical scale

(Toledo Weight Plate, Model I5S, Bay State, and Systems Inc. Burlington, MA). Body

mass index (BMI) was calculated as weight (kg) divided by height (m) squared [27, 28];

BMI was categorized into: recommended (BMI<25), overweight (25-30), and obese

(≥30) [114]. For the purpose of this analysis, vitamin D deficiency was defined as serum

25(OH)D concentration <50 nmol/L, insufficiency between 50-75 nmol/L, and

sufficiency as >75 nmol/L [54, 115]. Seasonality was defined as summer, July to

September; fall, October to December; winter, January to March; and spring, April to

June. Smoking status was categorized into yes or no for the current behavior. Vitamin D

from supplements was defined as: minimal 0-10 µg/day (200-400 IU/day); and high ≥ 10

µg/day (≥ 400 IU/day). Participant’s education was measured at baseline and defined by

the cutoff point of 8th grade, while poverty was measured at each time point and defined

by the U.S. Department of Health and Human Services federal poverty levels. Dietary

Vitamin D intake was estimated from a semi-quantitative food-frequency questionnaire

(FFQ) specifically designed for this population [91]. This FFQ has been validated against

serum carotenoids [116], and vitamin B12 [117] in a Hispanic population aged ≥ 60

years. Those with energy intakes < 600 or > 4800 kilocalories and/or > 10 questions

blank on the FFQ were excluded from the analyses. Nutrient intakes were calculated with

the Nutrition Data System for Research (NDS-R) software version 2007 (Nutrition

67

Coordinating Center, University of Minnesota, Minneapolis, MN, Food and Nutrient

Database 2007, released 2007). Alcohol intake (gm) was captured by the FFQ. Heart

disease status was self-reported. Medication classifications were based primarily on

AHFS Drug Information® (AHFS DI®) pharmacologic and therapeutic classification

2005. We only used cardiovascular pharmacologic medications based on the prefix 24.

Cardiovascular medication classification included: cardiac drugs, including glycoside

such as digoxin; antilipemic agents-HMG CoA reductase- such as Lipitor; omega-3 and -

6 supplements; hypotensive agents such as Thiazide; vasodilating-agents- including

nitrates, Viagra and Dypyridamole; alpha, beta and calcium channel blockers; ACE

inhibitors; angiotensin II inhibitors; and any hypertension medication [118].

Laboratory assays

Fasting blood samples (12-h) were drawn by a certified phlebotomist at the

participant’s home on the morning after the home interview, or as soon thereafter as

possible and kept at 4°C. Plasma was separated within 4 hours and was used for the

analysis of plasma lipid concentration. Aliquots of the isolated serum were saved and

stored at -80°C until processed. Serum 25(OH)D concentration was measured by

extraction followed by 25I radioimmunoassay Packard COBRA II Gamma Counter

(DiaSorin Inc., Stillwater, MN 55082 catalog # 68100E); intra- and inter-assay CV%

10.8% and 9.4%, respectively. Serum creatinine was assessed with a colorimetric, kinetic

reaction using the Olympus AU400e with Olympus Creatinine Reagents (OSR6178)

(Olympus America Inc., Melville, NY); intra-and inter-assay CV% were 2.0%, and 4.0%,

respectively. Plasma TC, HDL-C and TG were measured enzymatically on the Olympus

68

AU400. For TC, Olympus reagents (OSR6116) were used; intra- and inter-assay CV%

were 1.8%, and 2.2%, respectively. For HDL-C, Olympus reagents (OSR6156) were

used. For TG, Olympus reagents (OSR6033) were used; intra- and inter-assay CV% were

2.8%, and 2.7%, respectively. LDL-C and VLDL-C were calculated based on the

Friedewald equation from TC, HDL-C and TG i.e. VLDL=TG/5 and LDL-C =TC –

(VLDL + HDL-C) [119].

Statistical analyses:

All statistical analyses were conducted with SAS 9.1.3 (SAS Institute Inc., Cary,

NC); and all tests were 2-sided with P <0.05 to be considered statistically significant.

Dependent skewed variables were natural log-transformed to normalize their distributions

before the analysis. Demographic and clinical characteristics were compared using

Student’s t-test for continuous variables and the chi square tests for categorical variables.

Characteristics were compared by sex using two-tailed t-tests for two independent

samples, and χ2 test for differences in proportion. Paired t-tests were used to compare

characteristics across time points. Pearson partial linear correlation analysis was used to

examine relationships between serum 25(OH)D concentration with lipid profile at each

time point, adjusting for age and sex. Additionally, we constructed multivariate

regression analysis using general linear models to examine change in the plasma

concentration of the lipid profile components, with baseline serum 25(OH)D

concentration. A Risk factor model (RF-model) included: adjustment for age (y),

baseline-25(OH)D (nmol/L), BMI, baseline-lipid component, follow-up time (mos), and

sex in the full model. The Life-Style model (LS-model) included: physical activity, total

69

energy (kcal/day), current smoking status (y/n), alcohol (g/day), seasonality, poverty

level, and education in addition to the RF-model covariates. To control for confounding

by indications, the CVD model was used to adjust for the use of any cardiovascular

medication (y/n) as well as the prevalence of heart disease (y/n) [120]. For all

multivariate regression models, we checked assumptions of linearity and homogeneity by

examining outcome residuals against the exposure. For all final models, we performed

influence diagnostics tests to check for outliers and influential points. We also tested all

of our models for any potential effect modification by sex, with covariates in our models,

such as smoking, by including an interaction term with the exposure variable. Because

the lipid profile might be sex-specific, we constructed sex-stratified analyses in addition

to the full model. The baseline concentration of each lipid profile component was

included as a covariate to all of our models to control for regression to the mean [121].

To reduce bias and augment precision, we conducted multiple imputation analysis to

consider the uncertainty that missing values might convey within our dataset. We created

a total of five imputed data sets with (seed= 32, 37854) using procedure (PROC MI); we

included all covariates that are considered risk factors and being used in any of our

models. We used (PROC MIANALYZE) to combine all the five imputed data sets, thus,

producing valid statistical inferences that accounts for within and between imputation

variation [122]. We found that these missing values produced no significant impact on

our analysis.

70

RESULTS

Complete serum 25(OH)D concentration was available for 650 women and 255

men at both baseline and 2-yr. The mean ages of men and women were 59.5 ± 8.0 and

60.5 ± 8.0, respectively (Table 1). The mean follow-up time was 33.3± 12 months for

women and 30.5 ± 49 months for men. At baseline, the mean serum 25(OH)D of women

and men were, 44.0 ± 17.5 and 43.0 ± 16.1 nmol/L, respectively, while the 2-year values

were 50.1 ± 19.0 and 47.0 ± 17.5 nmol/L, respectively. The mean 25(OH)D at 2-yr was

5.5 nmol/L higher compared to baseline (P<0.0001) (Figure 1.1). Compared to men,

women had significantly higher plasma HDL-C, LDL-C and TC (mg/dL) at both time

points; while men had consistently higher plasma TG (mg/dL) and serum creatinine

(µmol/L). All lipid profile components increased significantly over time (P<0.0001). The

prevalence of reported heart disease status went up by 9 and 12% among men and

women, respectively. In addition, use of cardiovascular medication increased by 12% and

10.5% and of lipid lowering medication by 14% and 17%, for men and women,

respectively. While the mean BMI values did not differ across, the two time points,

women consistently had significantly higher mean BMI than men, (P<0.0001). In

contrast, waist circumference did not differ by sex, but was significantly higher at 2-yr

compared to baseline (P<0.0001).

There was no significant change over time in dietary vitamin D intake. However,

women had higher dietary vitamin D intake (IU/day) and higher vitamin D-containing

supplement use than men; while men had higher energy intake (kcal/day) than women at

both time points. There was a significant reduction in total energy consumed (P < 0.001)

over the 2.5 years for both men and women. The men walked significantly more

71

miles/day, i.e. 2 miles/day, compared to 1.4 miles/day (P<0.0001) for the women; men

also had higher physical activity scores and reported more current smoking and alcohol

consumption than women.

Among the lipid profile components, only HDL-C at 2-yr was significantly

correlated with serum 25(OH)D (r = 0.1, P =0.01) (Table 1.2). In longitudinal analysis

(Tables 1.3 and 1.4) baseline-25(OH)D was not associated with change in plasma HDL-C

concentration controlling for age, sex, BMI and follow-up time. However, in the CVD

model adjusting for physical activity, total energy (kcal/day), alcohol (gm/day), smoking

(y/n), seasonality, poverty level, education, heart disease (y/n) and taking CVD

medications (y/n), a significant association between baseline-25(OH)D and change in

plasma HDL-C concentration emerged (β = -0.04 ± 0.02, P =0.01), but in an unexpected

negative direction. Consistently, all three models explained about 62.5% of the variation

in the 2.5-year change in plasma HDL-C. Similarly, after differences in confounders were

controlled in LS-model, baseline-25(OH)D was significantly associated with change in

plasma TG, but in the unexpected positive direction (β = 0.40 ± 0.02, P =0.03). This

model explained only 30% of the variation in 2.5-year change in plasma TG. However,

when controlling for the CVD related covariates, the significant association was

attenuated (β = 0.34, P =0.05). Changes in plasma LDL-C and TC also tended to be

positively associated with baseline 25(OH)D , but these did not reach significance level

(P >0.05).

Smoking status was found to be an effect modifier of the association between

baseline-25(OH)D and HDL-C. The interaction between baseline-25(OH)D and smoking

status (P =0.04) was significant, thus, we stratified the analysis by smoking status for all

72

of the lipid profile components. Interestingly, baseline-25(OH)D was inversely yet

significantly associated with the change in plasma HDL-C concentration over time (β = -

0.06, P =0.002) only among non-smokers, when controlling for age, sex, BMI and

follow-up time. This association was strengthened after controlling for all life style

covariates (β = -0.06, P =0.0002) and CVD covariates (β = -0.07, P =0.0003). The

stratified models explained approximately 63 to 65% of the variation in 2.5-year change

in plasma HDL-C among non-smokers. On the other hand, null associations were found

among current smokers (Figure 1.2). Additionally, even though the interaction term

between baseline-25(OH)D and smoking status within the LDL-C models were not

significant, the stratified analysis showed that among currently smoking participants,

baseline-25(OH)D was inversely associated with the change in plasma LDL-C

concentration over time (β = -0.25, P =0.02), explaining roughly 42.3% of the model

variation. The addition of lifestyle covariates did not attenuate the coefficient, but the

significant association disappeared (β = -0.23, P =0.05) (Figure 1.3). Smoking status was

not an effect modifier of the association with baseline-25(OH)D and either plasma TG or

TC.

DISCUSSION

Although serum 25(OH)D concentration was positively associated with plasma

HDL-C cross-sectionally, longitudinal analyses showed inverse associations. Baseline-

25(OH)D demonstrated significant, negative association with 2.5-year change in plasma

HDL-C concentration (β = -0. 04, P =0.01), independent of important confounders

73

including age, sex, BMI, total energy intake, physical activity score, smoking, alcohol,

seasonality, poverty and education. This negative association remained after adjustment

for CVD-related variables. Similar relationships were seen with plasma TG, but not with

LDL-C and TC.

Our report is in contrast to several reports that indicate serum 25(OH)D

concentration is positively associated with plasma HDL-C, and negatively associated

with plasma LDL-C and TG [106, 107, 123-127]. In the prospective Health Professionals

Follow-up Study (HPFS), Giovannucci and colleagues demonstrated a positive linear

trend of LDL-C, HDL-C and TC, as well as a negative trend with plasma TG across

serum 25(OH)D quartiles among 900 US men, ages 64 and older [125]. Another study

showed that increasing serum 25(OH)D tertiles were independently associated with

higher HDL-C and lower TG; but not TC [107]. Interestingly, some the earlier studies

reported that both HDL-C and LDL-C were positively associated with serum 25(OH)D

concentration, thus, both increasing plasma concentration. It is well known that

increasing HDL-C and lowering LDL-C is desirable to improve CVD risk [2, 128]. A

cross-sectional analysis of the Third National Health and Nutrition Examination Survey

(NHANES III), found that serum 25(OH)D concentration was significantly inversely

associated with TC, in 13,331 US men and women ages 20 years and older [96]. The

study did not find an association between plasma HDL-C and serum 25(OH)D [96]. One

longitudinal study, with a 14 year follow-up in 1762 participants, found that serum

25(OH)D and plasma TG were inversely associated [104]. Most prior studies were cross-

sectional and no proof of a causal relationship can be inferred. In our longitudinal results,

with control for numerous confounders, we saw different results.

74

Our results are in accordance with a few reports that found that serum 25(OH)D

concentration may not be associated with an atheroprotective lipid profile [96, 104, 105,

123, 127]. One study found that TC was significantly associated with higher serum

25(OH)D concentration among 8,018 Norwegian men and women, aged 43-70 years

[104]. Two other studies also found that plasma TG were positively associated with

serum 25(OH)D; however, these latter studies were in children [123, 127].

Intervention studies in the form of placebo-controlled randomized trials (RCT) are

most informative for causal relationships. Very few RCT have been conducted to assess

the effect of vitamin D on lipid concentration, and most reported a null effect of vitamin

D supplementation on lipid profile [108, 129-131]. One study showed that daily vitamin

D supplementation (3320 IU/day) for one year significantly improved plasma TG: the

supplemented group lowered plasma TG compared to placebo, -13.5% and +3% (P

=0.001) correspondingly. However, the supplemented group also exhibited a significant

increase in plasma LDL-C [129]. Conversely, another study did not show that a year of

vitamin D supplementation, at 40, 000 IU/week, changed plasma lipids in 438

overweight/obese participants [108]. Yet another study showed that vitamin D

supplementation may not be atheroprotective in postmenopausal women [130]. Most of

these RCT can be criticized for low sample size, which may have contributed to the null

results.

Through stratification, we observed two distinct phenomena in the relationship

between plasma HDL-C, in particular, and serum 25(OH)D concentration. First, only

women retained the significant inverse association between serum 25(OH)D and the

change in plasma HDL-C concentration during follow up, thus, mirroring the association

75

of the combined data. The other explanation is that due to lack of power, we could not

detect a significant association of serum 25(OH)D and the change in plasma HDL-C

concentration among men. The significant association of serum 25(OH)D on HDL-C

among non-smokers was surprising, although similar results were reported previously

[132]. Among participants of the Tromsø study, smokers had 15-20% higher serum

25(OH)D concentration than non-smokers. The pathophysiological mechanism by which

smoking may impact serum 25(OH)D concentration or the metabolism of 1,25(OH)2D

has not been investigated. However, it has been suggested that smoking might influence

25-hydroxylase (CYP2R1) thus, lowering serum 25(OH)D concentration [133]. Another

point that might explain this paradox is that, smoking contains nicotine, which exerts

anti-inflammatory properties [134, 135].

Our results disagree with several epidemiologic and ecological studies regarding

the role of serum 25(OH)D on lipid profile as a major risk factor for CVD. One

explanation for our unexpected finding is that vitamin D induces the intestinal absorption

of dietary calcium, thus lowering calcium fecal excretion [37, 95]. This might affect the

amount of fatty acids in the feces, due to a reduction in the formation of insoluble fatty

soaps [136]. This could lead to an increased rate of the absorption of fatty acids, thus

higher plasma LDL-C and TG concentration, in particular, and TC in general [137]. In

addition, higher vitamin D could decrease cholesterol excretion because of the lower

availability of calcium within the intestine, and hence lower binding to bile acids, with

lower conversion of cholesterol to bile acids [138]. One study suggested that higher

calcium intake may lower LDL-C by increasing vitamin D A 15 week weight loss trial

showed that, among 63 healthy, overweight/obese women, the daily supplementation of

76

600 mg of calcium and 200 IU of vitamin D significantly improved LDL-C as well as

LDL: HDL-C ratio, while improvements in HDL-C and TG approached significance

[139].

Our findings should be interpreted within the context of a few limitations. First,

our population is Puerto Rican, therefore our findings could be very specific to this

particular admixture population, with genetic variation interacting not only with the

environment, but social factors as well [140-143]. This may limit generalizability.

Secondly, although we have controlled for many known potential covariates, residual

confounding is still be a possibility.

In conclusion, our longitudinal analysis showed that higher 25(OH)D serum

concentration did not translate into an increase of the atheroprotective plasma HDL-C,

nor a decrease of plasma TG in in this population of older Puerto Rican adults living in

the Boston area. Moreover, our results showed that smoking may be an effect modifier of

the longitudinal association of serum 25(OH)D and HDL-C, as well as LDL-C. The

conflicting information in the literature suggests that more studies are needed to clarify

the role of vitamin D on lipid profiles.

77

BIBLIOGRAPHY

1. Holick, M.F., The vitamin D deficiency pandemic and consequences for

nonskeletal health: mechanisms of action. Mol Aspects Med, 2008. 29(6): p. 361-8.

2. Holick, M.F., Vitamin D: extraskeletal health. Endocrinology and metabolism clinics of North America, 2010. 39(2): p. 381-400, table of contents.

3. Holick, M.F., High prevalence of vitamin D inadequacy and implications for

health. Mayo Clin Proc, 2006. 81(3): p. 353-73. 4. Holick, M.F., Resurrection of vitamin D deficiency and rickets. J Clin Invest,

2006. 116(8): p. 2062-72. 5. Holick, M.F., Vitamin D deficiency. N Engl J Med, 2007. 357(3): p. 266-81. 6. Holick, M.F., Deficiency of sunlight and vitamin D. Bmj, 2008. 336(7657): p.

1318-9. 7. Ramos-Lopez, E., et al., CYP2R1 (vitamin D 25-hydroxylase) gene is associated

with susceptibility to type 1 diabetes and vitamin D levels in Germans. Diabetes/Metabolism Research and Reviews, 2007. 23(8): p. 631-636.


vitamin D. Trends in biochemical sciences, 2004. 29(12): p. 664-73. 9. Cheng, J.B., et al., Genetic evidence that the human CYP2R1 enzyme is a key

vitamin D 25-hydroxylase. Proceedings of the National Academy of Sciences of the United States of America, 2004. 101(20): p. 7711-5.

10. Jones, G., Pharmacokinetics of vitamin D toxicity. The American Journal of Clinical Nutrition, 2008. 88(2): p. 582S-586S.

11. Deeb, K.K., D.L. Trump, and C.S. Johnson, Vitamin D signalling pathways in

cancer: potential for anticancer therapeutics. Nature reviews. Cancer, 2007. 7(9): p. 684-700.

12. Lacour, B., et al., Parathyroid function and lipid metabolism in the rat. Miner Electrolyte Metab, 1982. 7(3): p. 157-65.

13. Holick, M.F., Evolution and function of vitamin D. Recent Results Cancer Res, 2003. 164: p. 3-28.

14. Melamed, M.L., et al., 25-hydroxyvitamin d levels and the risk of mortality in the

general population. Archives of Internal Medicine, 2008. 168(15): p. 1629-1637. 15. Giovannucci, E., et al., 25-hydroxyvitamin D and risk of myocardial infarction in

men: a prospective study. Archives of Internal Medicine, 2008. 168(11): p. 1174-80.

16. Michos, E.D., et al., 25-Hydroxyvitamin D deficiency is associated with fatal

stroke among whites but not blacks: The NHANES-III linked mortality files. Nutrition, (0).

17. Ameri, P., et al., High prevalence of vitamin D deficiency and its association with

left ventricular dilation: an echocardiography study in elderly patients with

chronic heart failure. Nutrition Metabolism & Cardiovascular Diseases, 2010. 20(9): p. 633-40.

78

18. Freedman, D.M., et al., Serum Levels of Vitamin D Metabolites and Breast

Cancer Risk in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening

Trial. Cancer Epidemiology Biomarkers & Prevention, 2008. 17(4): p. 889-894. 19. Chen, T.C. and M.F. Holick, Vitamin D and prostate cancer prevention and

treatment. Trends Endocrinol Metab, 2003. 14(9): p. 423-30. 20. Wortsman, J., et al., Decreased bioavailability of vitamin D in obesity. The

American Journal of Clinical Nutrition, 2000. 72(3): p. 690-693. 21. Botella-Carretero, J.I., et al., Vitamin D deficiency is associated with the

metabolic syndrome in morbid obesity. Clinical Nutrition, 2007. 26(5): p. 573-80. 22. Kragt, J., et al., Higher levels of 25-hydroxyvitamin D are associated with a lower

incidence of multiple sclerosis only in women. Multiple Sclerosis, 2009. 15(1): p. 9-15.

23. Alvarez-Rodriguez, L., et al., Age and low levels of circulating vitamin D are

associated with impaired innate immune function. Journal of Leukocyte Biology, 2012. 91(5): p. 829-838.

24. Bellia, A., et al., Serum 25-hydroxyvitamin D levels are inversely associated with

systemic inflammation in severe obese subjects. Internal and Emergency Medicine, 2013. 8(1): p. 33-40.


Cardiovascular Health. The American Journal of Cardiology, 2012. 109(3): p. 359-363.

26. Pittas, A.G., et al., Systematic Review: Vitamin D and Cardiometabolic

Outcomes. Annals of Internal Medicine, 2010. 152(5): p. 307-314. 27. Sugden, J.A., et al., Vitamin D improves endothelial function in patients with Type

2 diabetes mellitus and low vitamin D levels. Diabet Med, 2008. 25(3): p. 320-5. 28. Zittermann, A., et al., Vitamin D supplementation enhances the beneficial effects

of weight loss on cardiovascular disease risk markers. American Journal of Clinical Nutrition, 2009. 89(5): p. 1321-7.

29. Kannel, W.B., et al., Serum Cholesterol, Lipoproteins, and the Risk of Coronary

Heart DiseaseThe Framingham Study. Annals of Internal Medicine, 1971. 74(1): p. 1-12.

30. Gordon, T., et al., High density lipoprotein as a protective factor against

coronary heart disease: The Framingham study. The American Journal of Medicine, 1977. 62(5): p. 707-714.

31. Hypponen, E., et al., 25-hydroxyvitamin D, IGF-1, and metabolic syndrome at 45

years of age: a cross-sectional study in the 1958 British Birth Cohort. Diabetes, 2008. 57(2): p. 298-305.

32. Jorde, R., et al., High serum 25-hydroxyvitamin D concentrations are associated

with a favorable serum lipid profile. European Journal of Clinical Nutrition, 2010. 64(12): p. 1457-64.

33. Karhapää, P., et al., Diverse associations of 25-hydroxyvitamin D and 1,25-

dihydroxy-vitamin D with dyslipidaemias. Journal of Internal Medicine, 2010. 268(6): p. 604-610.

79

34. Kazlauskaite, R., et al., Vitamin D is associated with atheroprotective high-

density lipoprotein profile in postmenopausal women. Journal of Clinical Lipidology, 2010. 4(2): p. 113-119.

35. Maki, K.C., et al., Serum 25-hydroxyvitamin D is independently associated with

high-density lipoprotein cholesterol and the metabolic syndrome in men and

women. Journal of Clinical Lipidology, 2009. 3(4): p. 289-296. 36. Jorde, R., et al., No improvement in cardiovascular risk factors in overweight and

obese subjects after supplementation with vitamin D3 for 1 year. Journal of Internal Medicine, 2010. 267(5): p. 462-472.

37. Mattei, J., et al., Allostatic load is associated with chronic conditions in the

Boston Puerto Rican Health Study. Social Science & Medicine, 2010. 70(12): p. 1988-96.

38. Tucker, K.L., et al., Self-reported prevalence and health correlates of functional

limitation among Massachusetts elderly Puerto Ricans, Dominicans, and non-

Hispanic white neighborhood comparison group. J Gerontol A Biol Sci Med Sci, 2000. 55(2): p. M90-7.

39. Tucker, K.L., O.I. Bermudez, and C. Castaneda, Type 2 diabetes is prevalent and

poorly controlled among Hispanic elders of Caribbean origin. Am J Public Health, 2000. 90(8): p. 1288-93.

40. Renzaho, A.M., J.A. Halliday, and C. Nowson, Vitamin D, obesity, and obesity-

related chronic disease among ethnic minorities: a systematic review. Nutrition, 2011. 27(9): p. 868-79.

41. Tucker, K., et al., The Boston Puerto Rican Health Study, a longitudinal cohort

study on health disparities in Puerto Rican adults: challenges and opportunities. BMC Public Health, 2010. 10(1): p. 107.

42. Bhupathiraju, S.N., et al., Centrally located body fat is associated with lower bone

mineral density in older Puerto Rican adults. The American Journal of Clinical Nutrition, 2011. 94(4): p. 1063-1070.

43. Flegal, K.M., et al., Overweight and obesity in the United States: prevalence and

trends, 1960-1994. International journal of obesity and related metabolic disorders : journal of the International Association for the Study of Obesity, 1998. 22(1): p. 39-47.

44. Holick, M.F., et al., Evaluation, Treatment, and Prevention of Vitamin D

Deficiency: an Endocrine Society Clinical Practice Guideline. Journal of Clinical Endocrinology & Metabolism, 2011. 96(7): p. 1911-1930.

45. Dawson-Hughes, B., et al., Estimates of optimal vitamin D status. Osteoporos Int, 2005. 16(7): p. 713-6.


Puerto Rican and non-Hispanic adults. American Journal of Epidemiology, 1998. 148(5): p. 507 - 518.

47. Bermudez, O., et al., Hispanic and non-Hispanic white elders from Massachusetts

have different patterns of carotenoid intake and plasma concentrations. Journal of Nutrition, 2005. 135(6): p. 1496 - 1502.

48. Kwan, L.L., O.I. Bermudez, and K.L. Tucker, Low vitamin B-12 intake and status

are more prevalent in Hispanic older adults of Caribbean origin than in

80

neighborhood-matched non-Hispanic whites. Journal of Nutrition, 2002. 132(7): p. 2059-64.

49. Pharmacists, A.S.o.H.-S., AHFS Pharmacologic-Therapeutic Classification 2008, American Society of Health-System Pharmacists.

50. Friedewald, W.T., R.I. Levy, and D.S. Fredrickson, Estimation of the

Concentration of Low-Density Lipoprotein Cholesterol in Plasma, Without Use of

the Preparative Ultracentrifuge. Clinical Chemistry, 1972. 18(6): p. 499-502. 51. Signorello, L.B., et al., Confounding by indication in epidemiologic studies of

commonly used analgesics. American Journal of Therapeutics, 2002. 9(3): p. 199-205.

52. Barnett, A.G., J.C. van der Pols, and A.J. Dobson, Regression to the mean: what it

is and how to deal with it. International Journal of Epidemiology, 2005. 34(1): p. 215-220.

53. Sterne, J.A.C., et al., Multiple imputation for missing data in epidemiological and

clinical research: potential and pitfalls. BMJ, 2009. 338. 54. Delvin, E.E., et al., Vitamin D Status Is Modestly Associated with Glycemia and

Indicators of Lipid Metabolism in French-Canadian Children and Adolescents. The Journal of Nutrition, 2010. 140(5): p. 987-991.

55. Forouhi, N.G., et al., Baseline Serum 25-Hydroxy Vitamin D Is Predictive of

Future Glycemic Status and Insulin Resistance: The Medical Research Council

Ely Prospective Study 1990–2000. Diabetes, 2008. 57(10): p. 2619-2625. 56. Giovannucci, E., et al., 25-hydroxyvitamin d and risk of myocardial infarction in

men: A prospective study. Archives of Internal Medicine, 2008. 168(11): p. 1174-1180.

57. Pilz, S., et al., Vitamin D and mortality in older men and women. Clinical Endocrinology, 2009. 71(5): p. 666-672.

58. Williams, D.M., A. Fraser, and D.A. Lawlor, Associations of vitamin D,

parathyroid hormone and calcium with cardiovascular risk factors in US

adolescents. Heart, 2011. 97(4): p. 315-320. 59. Bonow, R.O., Primary Prevention of Cardiovascular Disease. Circulation, 2002.

106(25): p. 3140-3141. 60. Balagopal, P., et al., Nontraditional Risk Factors and Biomarkers for

Cardiovascular Disease: Mechanistic, Research, and Clinical Considerations for

Youth. Circulation, 2011. 123(23): p. 2749-2769. 61. Zittermann, A., et al., Vitamin D supplementation enhances the beneficial effects

of weight loss on cardiovascular disease risk markers. The American Journal of Clinical Nutrition, 2009. 89(5): p. 1321-1327.

62. Heikkinen, A., et al., Long-term vitamin D3 supplementation may have adverse

effects on serum lipids during postmenopausal hormone replacement therapy. European Journal of Endocrinology, 1997. 137(5): p. 495-502.

63. Nagpal, J., J.N. Pande, and A. Bhartia, A double-blind, randomized, placebo-

controlled trial of the short-term effect of vitamin D3 supplementation on insulin

sensitivity in apparently healthy, middle-aged, centrally obese men. Diabetic Medicine, 2009. 26(1): p. 19-27.

81

64. Jorde, R., et al., Cross-sectional and longitudinal relation between serum 25-

hydroxyvitamin D and body mass index: the Tromsø study. European Journal of Nutrition, 2010. 49(7): p. 401-407.

65. Yoon, V., N.M. Maalouf, and K. Sakhaee, The effects of smoking on bone

metabolism. Osteoporosis International, 2012. 23(8): p. 2081-2092. 66. Reid, I.R., Effects of calcium supplementation on circulating lipids: potential

pharmacoeconomic implications. Drugs Aging, 2004. 21(1): p. 7-17. 67. Hu, F.B., J.E. Manson, and W.C. Willett, Types of Dietary Fat and Risk of

Coronary Heart Disease: A Critical Review. Journal of the American College of Nutrition, 2001. 20(1): p. 5-19.

68. Van der Meer, R., et al., Effects of supplemental dietary calcium on the intestinal

association of calcium, phosphate, and bile acids. Gastroenterology, 1990. 99(6): p. 1653-9.

69. Major, G.C., et al., Supplementation with calcium + vitamin D enhances the

beneficial effect of weight loss on plasma lipid and lipoprotein concentrations. American Journal of Clinical Nutrition, 2007. 85(1): p. 54-9.

70. Junyent, M., et al., The effects of ABCG5/G8 polymorphisms on plasma HDL

cholesterol concentrations depend on smoking habit in the Boston Puerto Rican

Health Study. Journal of Lipid Research, 2009. 50(3): p. 565-73. 71. Lai, C.-Q., et al., Population admixture associated with disease prevalence in the

Boston Puerto Rican health study. Human Genetics, 2009. 125(2): p. 199-209. 72. Mattei, J., et al., Apolipoprotein A5 polymorphisms interact with total dietary fat

intake in association with markers of metabolic syndrome in Puerto Rican older

adults. Journal of Nutrition, 2009. 139(12): p. 2301-8. 73. Mattei, J., et al., Disparities in allele frequencies and population differentiation

for 101 disease-associated single nucleotide polymorphisms between Puerto

Ricans and non-Hispanic whites. BMC Genetics, 2009. 14(10): p. 45.

82

TABLE 1.1: Descriptive characteristic of men and women in the Boston Puerto Rican Health Study (BPRHS) & Boston Puerto Rican Osteoporosis Study (BPROS)

Baseline (n=1504) Two Year (n=970 )

Women (n=1060)

Men (n=444)

Women (n=700)

Men (n=270)

Age (y) 57.0 ± 7.3 56.0 ± 8.0 60.3 ± 7.5 59.5 ± 8.0

Education (>8th grade %) 51.5 55.5 -- --

Poverty (%) 60.0 49.0$$ 59.0 48.5$$

25(OH)D (nmol/L) §§§ 44.0± 17.4 43.0 ± 16.5 50.1 ± 19.0 47.0 ± 17.5*

Follow-up time (months) -- -- 33.3± 12 30.5± 49

Serum Lipids

HDL (mg/dL) §§§ 47.1 ± 12.2 40.5 ± 12.5*** 48.1 ± 12.5 41.2 ± 12.0***

LDL (mg/dL) §§§ 110 ± 34.4 101 ± 35.2*** 112 ± 35.3 103 ± 34.5**

Triglyceride (mg/dL) §§ 157 ± 99.0 175 ± 140** 154 ± 78.5 178± 151*

Total cholestrol (mg/dL) §§§ 188 ± 40.5 174 ± 43.0*** 190 ± 42.0 177± 42.2***

Creatinine (µmol/L) §§§ 68.0 ± 27.5 88.0 ± 45.0*** 69.4 ± 20.2 88.6 ± 29.4***

Heart disease (%) 20.5 22.0 23.0 24.0

Cardiovascular meds (%) 61.5 57.3 68.0 64.3

Lipid-lowering meds (%) 41.0 38.5 48.0 44.0

Adiposity Measures

BMI (kg/m2) 33.2 ± 7.0 30.0 ± 5.3*** 33.0 ± 7.0 30.2 ± 5.5***

Waist Circumference(cm) §§§ 102.2 ± 15.4 103.0 ± 13.5 104.0 ± 15.0 104.3 ± 14.3

Other Covariates

83

Miles walking/day§§§ 1.0 ± 1.5 1.4 ± 2.0*** 1.4 ± 2.0 2.0 ± 2.4 ***

Physical Activity Score # 31.0 ±1.13 32.1±1.20*** 31.3 ± 1.1 32.5 ± 1.2***

Smoking (%) 20.1 32.5$$$ 19.0 29.0$$

Alcohol drinking (%) 36.0 52.0$$$ 28.0 43.0$$$

Alcohol intake (gm/day) ** 3.0 ± 13.5 14.0 ± 45.0$$$ 2.5 ±17.0 7.0 ± 24.0$$$

Vitamin D supp (>200IU/day) (%)

30.0 24.1 30.2 18.5$$

Total vitamin D (IU/day) 290.5±172.0 280.0± 178.0 287.5±152.0 257.0±123.0**

Total Energy (kcal/day) §§§ 2100 ± 1100 2750 ± 1430*** 1750 ±865 2215 ±1120***

BMI, body mass index; 25(OH)D, 25-hydroxyvitamin D. # Geometric mean ± SD 1Data are presented as mean ± SD for continuous variables or as % for categorical variables Means were compared using t-test for between sex (within each time points): *P -value <0.05, ** P -value <0.001, ***P-value <0.0001

Means were compared using paired t-test for between time points: §P -value <0.05, §§P -

value <0.001, §§§P-value <0.0001

Percentages were compared using Chi-square test for between sex (within each time points): $P -value <0.05, $$

P -value <0.001, $$$ P-value <0.0001

84

TABLE 1.2: Pearson partial linear correlation of serum 25(OH)D concentration and

adiposity phenotype at baseline and 2-year follow-up

Lipid Profile

Components

(mg/dL)

Baseline Two-Year

25(OH)D p-value 25(OH)D p-value

HDL-C 0.01 0.7 0.1 0.01

LDL-C 0.01 0.7 -0.02 0.4

TC -0.03 0.3 -0.03 0.3

TG 0.005 0.8 -0.01 0.7

Pearson partial linear correlation coefficient (r) is age, sex and seasonality adjusted Abbreviations: high-density lipoprotein (HDL-C); low-density lipoprotein (LDL-C); triglycerides (TG); total cholesterol (TC); 25-hydroxy vitamin D concentration (25(OH)D)

85

TABLE 1.3 Regression coefficients (± SE) from longitudinal linear regression analysis of lipid profile components and baseline

serum 25(OH)D concentration

Lipid Profile Components

Baseline 25(OH)D

(nmol/L)

Change in HDL-C Change in LDL-C

ALL

(n=866)

Men

(n=246)

Women

(n=620)

ALL

(n=864)

Men

(n=245)

Women

(n=619)

β ± SE p β ±SE p β ± SE p β ± SE p β ±SE p β ± SE p

Risk Factor model -0.025

± 0.01

0.1

-0.01

±0.02

0.6

- 0.03

± 0.01

0.1 -0.05

±0.05

0.3

-0.20

± 0.1

0.07

-0.003

±0.1

0.9

Life-Style

model

-0.045

±0.01

0.007

-0.04

± 0.03

0.2

- 0.05

± 0.02

0.01 -0.12

±0.06

0.05

-0.22

±0.12

0.07

-0.1

± 0.1

0.2

CVD Model -0.04

±0.02 0.01

-0.035

±0.03 0.2

- 0.043

± 0.02 0.02

-0.01

±0.05 0.09

-0.2 ±

0.1 0.1

-0.1

± 0.1 0.2

Risk factor model: adjusted for age, BMI, baseline-lipid-component (mg/dL), follow-up time (mos), and sex, in the case of the combined data1

86

Life-Style model: adjusted for model 1 variables physical activity, total energy (kcal/day), alcohol, smoking, seasonality, poverty level, education and sex in the case of the combined data1

CVD model: adjusted for model 1+ 2 variables plus heart-disease (y/n), taking cardiovascular agents (y/n) and sex in the case of the combined data1

87

TABLE 1.4 Regression coefficients (± SE) from longitudinal linear regression analysis of lipid profile components and baseline

serum 25(OH)D concentration

Lipid Profile Components

Baseline 25(OH)D

(nmol/L)

Change in TG Change in TC

ALL

(n=844)

Men

(n=228)

Women

(n=615)

ALL

(n=844)

Men

(n=228)

Women

(n=615)


Risk Factor model 0.22

± 0.2

0.2

0.3

± 0.5

0.6

0.2

± 0.15

0.2 -0.04 ±

0.06

0.5

-0.11

± 0.1

0.4

-0.02

± 0.1

0.7

Life-Style

model

0.4

± 0.2

0.03

1.0

± 0.5

0.1

0.2

± 0.17

0.2 - 0.12 ±

0.1

0.08

-0.1

± 0.1

0.5

- 0.14

± 0.1

0.08

CVD Model 0.34

± 0.2 0.05

1.0

± 0.5 0.08

0.16

± 0.15 0.3

-0.11

± 0.1 0.1

-0.05

± 0.1 0.7

-0.13

± 0.1 0.09

Risk factor model: adjusted for age, BMI, baseline-lipid-component, time f/up (mos), and sex in the case of the combined data1 Life-Style model: adjusted for model 1 variables physical activity, total energy (kcal/day), alcohol, smoking, seasonality, poverty level, education and sex in the case of the combined data1

88

CVD model: adjusted for model 1 variables plus heart-disease (y/n), taking cardiovascular agents (y/n) and sex in the case of the combined data1.

89

FIGURE 1.1 Prevalence of low serum 25-hydroxyvitamin D concentration among Puerto Rican men (n=254) and women (n=667). Serum 25(OH)D concentration was categorized as: severe deficiency ≤ 25 nmol/l; deficiency 25-50 nmol/l, insufficiency 50-75 nmol/L, and sufficiency ≥ 75 nmol/L. Values are percentage of individuals.

0

10

20

30

40

50

60

Baseline 2-YR Baseline 2-YR

PE

RC

EN

TA

GE

≤ 25 nmol/L

25- 50 nmol/L

50 - 75 nmol/L

≥75 nmol/L

Men (n=254) Women (n=667)

90

FIGURE 1.2 The association between serum (25)OH at baseline and the change in HDL a (mg/dL) over time modified by smoking . M1: adjusted for age, BMI, baseline-lipid-component (mg/dL), time f/up (mos), and sex. M2: adjusted for model 1 variables physical activity, total energy (kcal/day), alcohol, seasonality, poverty level, education. M3: adjusted for model 1 + 2 variables plus heart-disease (y/n), taking cardiovascular agents (y/n)

91

FIGURE 1.3 The association between serum (25)OH at baseline and the change in LDL among

others over time modified by smoking. M1: adjusted for age, BMI, baseline-lipid-component (mg/dL), time f/up (mos), and sex. M2: adjusted for model 1 variables plus physical activity score, total energy (kcal/day), alcohol, seasonality, poverty level, education M3: adjusted for model 1+ 2 variables plus heart-disease (y/n), taking cardiovascular agents (y/n)

92

CHAPTER THREE

An Inverse Association of Vitamin D Serum Concentration and Inflammatory

Markers

The Boston Puerto Rican Osteoporosis Study


1Department of Pharmaceutical Science Department. 3Department of Health Sciences, 2Bouvé

College of Health Sciences. Northeastern University.


Presented as a Poster at Experimental Biology 2013

93

ABSTRACT

Vitamin D deficiency has been linked to obesity, metabolic syndrome and autoimmune

diseases. Several reports suggest an association between serum 25-hydroxy vitamin D

concentration (25(OH)D) and inflammatory markers. We cross-sectionally examined the

relation between serum 25(OH)D concentration and serum inflammatory markers in 963 Puerto

Rican adults: 268 men and 695 women, aged 45-75 years. After adjusting for age, sex, and

seasonality, serum 25(OH)D was significantly negatively associated with tumor necrosis factor-

alpha (TNF-α) and Interleukin 6 (IL-6) (P<0.01 and P<0.04, respectively), but not with high-

sensitivity C-reactive protein (hsCRP). These associations remained significant adjusting for

waist circumference (WC), total body fat (TBF), and abdominal fat mass; and for smoking

status, alcohol intake, vitamin D supplementation and physical activity. In conclusion, lower

serum 25(OH)D concentration was significantly associated with higher concentration of

inflammatory markers in this population of older Puerto Rican adults living in the Boston area.

Longitudinal studies are needed to clarify causal pathways.

94

Abbreviations:

25-hydroxyvitamin D concentration (25(OH)D); android fat-mass (AFM); cardiovascular disease

(CVD); confidence interval [CI]; dual-energy x-ray absorptiometry (DXA); gynoid fat-mass

(GFM); high-sensitivity C-reactive protein (hsCRP); Human tumor necrosis factor-alpha (TNF-

α); Interleukin 6 (IL-6); Life-style model (LS-model); Non-steroidal anti-inflammatory drugs

(NSAIDs); Obesity model (O-model); Intact parathyroid hormone (iPTH); percent body fat

(%TBF); Risk factor model (RF-model); total body fat (TBF); white blood cell (WBC); food-

frequency questionnaire (FFQ).

95

INTRODUCTION

It is well recognized that vitamin D actions extend beyond calcium metabolism and bone

health [93, 94]. As a steroid hormone, vitamin D functions as a gene modulator for several

hundred genes. It regulates both-directly and indirectly the proliferation and differentiation of

immune cells including macrophages, monocytes and lymphocytes; initiating a phosphorylation

cascade of mitogen-activated protein kinases (MAPK) and extracellular signal-regulated kinase

(ERK); potentially participating in a cross-talk with the genomic pathway to modulate gene

expression [37-39, 144]. The expression of VDR in various types of cells, including immune

cells, supports the assumption that vitamin D may exert some immune modulator actions [4, 145,

146]. To date, a wide range of immune responses have been reported for vitamin D, including

the down regulation of a number of inflammatory responses, such as Interleukin 6 (IL-6), Human

tumor necrosis factor-alpha (TNF-α); and interferon-γ (IFN- γ) [4, 72, 145, 147] as well as high-

sensitivity C-reactive protein (hsCRP), a non-specific inflammatory marker [148, 149].

However, most of these reports are from animal studies, along with some in vitro work [147,

150-152]. Few have investigated the association of vitamin D and inflammation in humans [98,

153-155]. One observational study showed that serum 25(OH)D concentration was inversely

associated with TNF-α in women, after controlling for potential covariates [154]. Another study

showed that not only TNF-α, but also IL-6 and hsCRP, were inversely associated with 25(OH)D

[98]. In a double-blind, randomized, placebo-controlled trial (RCT), nine months of daily

supplementation with 2000 IU of vitamin D and 500 mg of calcium resulted in an improved

cytokine profile; in particular, lower TNF-α concentration were seen compared to baseline [153].

In a large randomized intervention trial, Beilfuss and colleagues demonstrated one year of

supplementation with vitamin D in the range between 40,000 to 20,000 IU/week, led to

96

marginally significant lower serum IL-6 concentration in the intervention arm compared to the

placebo arm. Interestingly, however, they observed significantly elevated hsCRP concentration

within the intervention arm [155].

Ethnic minorities generally, and Puerto Ricans specifically, carry a higher burden of

health disparities compared to non-Hispanic whites; which translates to a greater risk of chronic

disease [109-111]. Several studies have reported that ethnic minorities tend to have higher

concentration of inflammatory marker [156]. One systematic review showed that ethnic

minorities were considerably more likely to have vitamin D deficiency, across all age groups.

They concluded a number of potential mechanisms link vitamin D deficiency to obesity-related

chronic disease, including inflammation [112].

The aim of this study was to determine the cross-sectional association between serum

25(OH)D concentration and inflammatory markers, namely, IL-6, TNF-α, and hsCRP, among

participants of the Boston Puerto Rican Osteoporosis Study (BPROS).



Recruitment and data collection for the BPROS are described in detail elsewhere [113].

In brief, a bilingual interviewer conducted home visits and administrated a questionnaire that

collects information on socioeconomic status, education, household income, health and health

behaviors, medical diagnoses, current medications, depression, and psychological acculturation.

The interviewer measured anthropometrics, physical performance and blood pressure. The

following day, the study phlebotomist collected 24-hr urine, 12-hr fasting blood and saliva.

97

Participants were provided with written instructions for collecting urine, taking PABA tablets,

and fasting for the blood draw. Two days before the scheduled visit, participants were given a

reminder phone call and reinforced instruction for urine collection and overnight fasting.

Participants were asked to provide the 24-hour urine collection in two containers, one for every

12 hours. Once the 2-yr follow up visit was complete, participants were re-consented for the

BPROS. They visited the Human Nutrition Research Center on Aging (HNRCA) at Tufts

University for bone density and body composition measurements, and an additional blood draw.

Participants then completed additional questionnaires on sun exposure, family history of

osteoporosis, and history of fracture. All laboratory measurements for this study were performed

by the Nutrition Evaluation Laboratory at Tufts University.

By May 2013, 1257 participants completed 2year follow-up visits and 970 completed the

BPROS study. All study protocols were approved by the Institutional Review Boards at Tufts

Medical Center and Northeastern University, and each participant provided written informed

consent. There were no significant differences in socio-demographic characteristics between

those with complete or incomplete information. The present study includes complete data for 700

Puerto Rican women and 270 Puerto Rican men.

Covariates and Exposure Assessments

At each time point, in addition to the designed questionnaires, physical function tests

were completed and anthropometric measures were taken. w. Weight, and waist and hip

circumferences were measured in duplicate. Weight was measured using a clinical scale (Toledo

Weight Plate, Model I5S, Bay State, and Systems Inc. Burlington, MA). Body mass index (BMI)

98

was calculated as weight (kg) divided by height (m) squared [84, 85]. Diastolic and systolic

blood pressure were measured with an electronic sphygmomanometer (Dinamap TM Model

8260, Critikon, Tampa, FL), at three time points during home visits by trained interviewers: once

near the beginning, the middle, and then near the end of the interview, after a short rest during

which the subject was seated quietly. The average of the second and third readings was used in

the analysis.

Education was measured at baseline and defined using the 8th grade as a cutoff point;

poverty was measured at each time point and defined by the U.S. Department of Health and

Human Services federal poverty levels. Dietary vitamin D intake was estimated from a semi-

quantitative food-frequency questionnaire (FFQ) that was specifically designed for this

population [91]. This FFQ has been validated against serum carotenoids [116], and vitamin B12

[117] in a Hispanic population aged ≥ 60 years. Those with energy intakes < 600 or > 4800

kilocalories and/or > 10 questions blank on the FFQ were excluded from the analyses. Nutrient

intakes were calculated with the Nutrition Data System for Research (NDS-R) software version

2007 (Nutrition Coordinating Center, University of Minnesota, Minneapolis, MN, Food and

Nutrient Database 2007, released 2007). Alcohol intake (gm) was captured by the FFQ. For the

purpose of this paper, vitamin D deficiency was defined as serum 25(OH)D concentration below

50 nmol/L, insufficiency between 50-75 nmol/L, and sufficiency above 75 nmol/L [26, 157].

Seasonality was defined as summer, July to September; fall, October to December; winter,

January to March; and spring, April to June. Smoking status was categorized as yes or no for the

current behavior. Vitamin D from supplements was defined as: minimal 0-10 µg/day (200-400

IU/day); and high ≥ 10 µg/day (≥ 400 IU/day).

99

Laboratory assays

Fasting blood samples (12-h) were drawn by a certified phlebotomist at the participant’s

home on the morning after the home interview, or as soon thereafter as possible. Aliquots were

saved and stored at -80°C until processed. Serum 25(OH)D concentration was measured by

extraction, followed by 25I radioimmunoassay Packard COBRA II Gamma Counter (DiaSorin

Inc., Stillwater, MN 55082 catalog # 68100E); intra- and inter-assay CV% were 10.8% and

9.4%, respectively. Intact serum parathyroid hormone (IPTH) was detected by two-site

chemiluminescent enzyme-labeled immunometric assay (Diagnostic Products Corporation

(DCP) Los Angeles, CA ) [158]; intra- and inter-assay CV% were 5.5%, and 7.9%, respectively.

Serum creatinine was assessed with a colorimetric, kinetic reaction using the Olympus AU400e

with Olympus Creatinine Reagents (OSR6178) (Olympus America Inc., Melville, NY); intra-

and inter-assay CV% were 2.0%, and 4.0%, respectively. The hsCRP was measured by solid-

phase chemiluminescent immunometric assay; intra- and inter-assay CV% were 6.0%, and 7.3%,

respectively. Both IL-6 was and TNF-α were measured using commercial radioimmunoassay kit

(R&D Systems Inc, Minneapolis, MN). For IL-6, the intra- and inter-assay CV% was 1.6-4.2 %

and 3.3-6.4 %, respectively. For TNF-α, the intra- and inter-assay CV% was 5.3-8.8 %% and

10.8-16.7 % respectively. WBC (mm3) was measured by electronic impedance using ABX

Pentra 60 C+ (Baker Instruments Corp., Allentown, PA)


All statistical analyses were conducted with SAS 9.1.3 (SAS Institute Inc, Cary, NC); and

all tests were 2-sided, with P <0.05 considered statistically significant. Dependent skewed

variables were natural log-transformed to normalize their distributions before the analysis.

100

Demographic and clinical characteristics were compared using Student’s t-tests for continuous

variables, and chi square tests for categorical variables. Characteristics were compared across sex

using two-tailed t-tests for two independent samples, and χ2 test for differences in proportion. We

applied Pearson Linear correlations to examine relationships between serum 25(OH)D

concentration and the different inflammatory markers, adjusting for age and sex. Further, we

applied the general linear models procedures (PROC GLM) to model the cross-sectional

associations between 25(OH)D (continuous) and inflammatory markers (continuous), and

adjusted for potential confounding variables. Linear trends of the inflammatory variables across

tertiles of 25(OH)D were adjusted for sex, seasonality, and age as well as other potential

covariates. The multivariate regression models for both IL-6 and TNF-α are: Risk-Factor model

(RF-model): adjusted for age (y), sex, seasonality, WBC count (mm3), and use of non-steroidal

anti-inflammatory (Y/N); the Obesity model (O-model): serum iPTH (pg/ml), BMI, total-fat

mass (g), and abdominal fat-mass (AFM), in addition to the prior model’s covariates; the Life-

Style model (LS-model): adjusted for the prior two models, in addition to physical activity,

current smoking status (y/n), alcohol intake (g/d), poverty level, and education. Some models

were adjusted for further covariates. The multivariate regression models for hsCRP are: Risk-

Factor model (RF-model): Sex, age seasonality, the use of non-steroidal anti-inflammatory

medication (Y/N), energy intake (kcal/day), smoking (Y/N), and WBC count (mm3) (to control

for current infection). Obesity model (O-model): RF-model+ BMI, total fat-mass (g), diastolic

blood pressure and diabetes (Y/N). Comprehensive Model (C-model): RF-model+ O-model,

poverty, and education. Linear trend tests were performed by assigning each participant the

median 25(OH)D for each tertile category, and using it as a continuous variable in PROC GLM.

For all linear models, we checked the assumptions of linearity and homogeneity through

101

examining the outcome residuals against the exposure. For all final models, we performed

influence diagnostics tests to check for outliers and influential points.

RESULTS

The mean ages of women and men were 60.3 ± 7.5 and 59.5 ± 8.0, respectively (Table

2.1). The mean serum 25(OH)D concentration for women was 50.1 ± 19.0, and was significantly

higher (P = 0.01) than for men, 47.0 ± 17.5 nmol/L. There was no significant difference between

men and women for most inflammatory markers, with the exception of hsCRP; where women

had significantly higher concentration than men, 3.5± 3.3 and 2.3±3.2 mg/L, respectively.

Women also had significantly higher adiposity measures than men, with the exception of WC,

where no significant difference was found (although this implies higher WC for women relative

to recommended WC). Puerto Rican women had significantly higher dietary vitamin D and

supplement intakes, and lower total energy intake (kcal/day) than men. Puerto Rican men walked

significantly more miles/day and had higher physical activity score compared to women. More

men reporting current smoking and current alcohol consumption compared to women, while

more women used non-steroidal anti-inflammatory drugs (NSAID) than did men 34.0% vs.

18.0%, respectively.

Both IL-6 and TNF-α concentrations showed significant inverse associations with

25(OH)D, (r = -0.1, P =0.0005) and (r = -0.11, P =0.001), respectively (Table 2.3). Per unit

increase in 25(OH)D, there was a corresponding significant reduction in IL-6 (β= -0.03 ±0.001,

P =0.01) and TNF-α (β=-0.002 ± 0.001, P =0.02). Further adjustment for covariates of both the

102

obesity and the life-style models did not attenuate these significant inverse associations.

Conversely, hsCRP showed no significant association with 25(OH)D in any model (Table 2.4).

We further evaluated the adjusted means of each inflammatory marker across the tertiles

of serum 25(OH)D concentration. Participants in the second and third tertile of 25(OH)D had

significantly lower IL-6 concentration compared to those in the lowest tertile, < 40 nmol/L, P

=0.04 and P =0.003, respectively (Figure 2.1). A similar trend was observed for TNF-α; those in

the highest tertile of 25(OH)D had significantly lower TNF-α concentration compared to those in

the lowest tertile of 25(OH)D, < 40 nmol/L (P =0.007) (Figure 2.2). There were significant

decreasing trends in IL-6 and TNF-α (P-trend = 0.002 and P-trend = 0.001) across 25(OH)D

tertiles. No significant trend was observed for hsCRP concentration (Figure 2.3)

DISCUSSION

The present study provides evidence of an inverse association between serum 25(OH)D

and inflammatory markers, namely, IL-6 and TNF-α, among participants of the BPROS;

irrespective of age, sex, seasonality, adiposity phenotype and lifestyle factors that include

smoking, physical activity and alcohol consumption. This population, as reported before,

exhibits a high prevalence of vitamin D deficiency [159]. After adjustment for potential

covariates, serum 25(OH)D concentration remained significantly associated with IL-6 and TNF-

α, but was not associated with hsCRP., These results are in accordance with several other reports

on the association between lower serum 25(OH)D concentration and inflammatory markers [98,

154, 155]. These results are also consistent with in vitro and animal studies that showed TNF-α

suppression by vitamin D [97, 152, 160, 161] and IL-6 [97, 147, 162]. One RCT, in heart failure

103

patients, demonstrated that 9 months of daily supplementation with 2000 IU of vitamin D and

500 mg of calcium significantly lowered serum TNF-α concentration [153]. Another RCT also

demonstrated a suppression of TNF-α concentration with a daily dose of a preparation that

contained 3332 IU of vitamin D [129].

The lack of association between serum 25(OH)D and hsCRP is not in agreement with a

number of cross-sectional reports [98, 163]. Possible explanations include that, first, hsCRP

concentration can be affected by acute phase reactions due to viral infection, which may increase

hsCRP by 50,000 fold. However, this is unlikely to be the case with the presence of constant yet

potent stimuli [149]. Another point to emphasize is that this specific population has a high

prevalence of health disparities that include metabolic syndrome, depression, and diabetes, which

might mask the association [109, 164]. Additionally, in this population the mean of 25(OH)D

concentration is low, with widespread vitamin D deficiency and thus, limited distribution and a

very low number of sufficient levels to show significant differences. In our analysis, the

adjustment of white blood cell counts, which is an indicator of current infection, did not improve

the null association between hsCRP and serum 25(OH)D concentration.

Higher serum concentration of inflammatory markers such TNF-α, IL-6 and hsCRP have

emerged as independent risk factors for CVD, diabetes and several autoimmune diseases, and are

often correlated with clinical impairment [165, 166]. Therefore, attenuating inflammatory marker

concentration may reduce risk of these conditions. When bound to its receptor, 1,25(OH)2D, the

active metabolite, has the ability to attenuate the high levels of inflammatory markers by down-

regulating a number of related genes [37-39, 144]. Our data suggest that serum 25(OH)D

concentration explains about 6-20% of the variation in the IL-6 and TNF-α concentration,

respectively, among our participants.

104

The findings of this report need to be understood in the context of some limitations. As a

cross-sectional study, it cannot provide further insight about causation or long-term status.

Longitudinal studies with two or more time points are needed to investigate the role of serum

25(OH)D concentration on inflammatory response. Although we were able to adjust for a

number of known cofounders, residual confounding might still a possibility.

In conclusion, the present study supports the hypothesis that inflammatory markers such

as IL-6 and TNF-α are inversely associated with serum 25(OH)D concentration in Puerto Rican

adults. These findings support the need to re-examine the determination of optimal vitamin D

status. However, more studies, especially longitudinal prospective studies, are needed to fully

characterize the relationship between serum 25(OH)D concentration and inflammatory markers.

105

BIBLIOGRAPHY

1. Holick, M.F., The vitamin D deficiency pandemic and consequences for nonskeletal

health: mechanisms of action. Mol Aspects Med, 2008. 29(6): p. 361-8. 2. Holick, M.F., Vitamin D: extraskeletal health. Endocrinology and metabolism clinics of

North America, 2010. 39(2): p. 381-400, table of contents. 3. Plum, L.A. and H.F. DeLuca, Vitamin D, disease and therapeutic opportunities. Nat Rev

Drug Discov, 2010. 9(12): p. 941-55. 4. Deeb, K.K., D.L. Trump, and C.S. Johnson, Vitamin D signalling pathways in cancer:

potential for anticancer therapeutics. Nature reviews. Cancer, 2007. 7(9): p. 684-700. 5. POLLY, P., et al., VDR-Alien: a novel, DNA-selective vitamin D3 receptor-corepressor

partnership. The FASEB Journal, 2000. 14(10): p. 1455-1463. 6. 2010 ICD-9-CM.

7. Guillot, X., et al., Vitamin D and inflammation. Joint Bone Spine, 2010. 77(6): p. 552-557.

8. Krishnan, A.V. and D. Feldman, Mechanisms of the anti-cancer and anti-inflammatory

actions of vitamin D. Annual review of pharmacology and toxicology, 2011. 51: p. 311-36.

9. Querfeld, U., Vitamin D and inflammation. Pediatr Nephrol, 2013. 28(4): p. 605-10. 10. Dickie, L.J., et al., Vitamin D3 down-regulates intracellular Toll-like receptor 9

expression and Toll-like receptor 9-induced IL-6 production in human monocytes. Rheumatology (Oxford), 2010. 49(8): p. 1466-1471.

11. Jacques M, L., Immunomodulatory actions of 1,25-Dihydroxyvitamin D3. The Journal of Steroid Biochemistry and Molecular Biology, 1995. 53(1–6): p. 599-602.

12. Wilson, A.M., M.C. Ryan, and A.J. Boyle, The novel role of C-reactive protein in

cardiovascular disease: risk marker or pathogen. Int J Cardiol, 2006. 106(3): p. 291-7. 13. Gabay, C. and I. Kushner, Acute-phase proteins and other systemic responses to

inflammation. N Engl J Med, 1999. 340(6): p. 448-54. 14. Rigby, W.F., S. Denome, and M.W. Fanger, Regulation of lymphokine production and

human T lymphocyte activation by 1,25-dihydroxyvitamin D3. Specific inhibition at the

level of messenger RNA. The Journal of Clinical Investigation, 1987. 79(6): p. 1659-1664.

15. Rigby, W.F., T. Stacy, and M.W. Fanger, Inhibition of T lymphocyte mitogenesis by 1,25-

dihydroxyvitamin D3 (calcitriol). The Journal of Clinical Investigation, 1984. 74(4): p. 1451-1455.

16. Zhu, Y., et al., Calcium and 1α,25-dihydroxyvitamin D3 target the TNF-α pathway to

suppress experimental inflammatory bowel disease. European Journal of Immunology, 2005. 35(1): p. 217-224.

17. Schleithoff, S.S., et al., Vitamin D supplementation improves cytokine profiles in patients

with congestive heart failure: a double-blind, randomized, placebo-controlled trial. American Journal of Clinical Nutrition, 2006. 83(4): p. 754-9.

18. Peterson, C. and M. Heffernan, Serum tumor necrosis factor-alpha concentrations are

negatively correlated with serum 25(OH)D concentrations in healthy women. Journal of Inflammation, 2008. 5(1): p. 10.

106

19. Beilfuss, J., et al., Effects of a 1-year supplementation with cholecalciferol on interleukin-

6, tumor necrosis factor-alpha and insulin resistance in overweight and obese subjects. Cytokine, 2012. 60(3): p. 870-4.



21. Mattei, J., et al., Allostatic load is associated with chronic conditions in the Boston

Puerto Rican Health Study. Social Science & Medicine, 2010. 70(12): p. 1988-96. 22. Tucker, K.L., et al., Self-reported prevalence and health correlates of functional

limitation among Massachusetts elderly Puerto Ricans, Dominicans, and non-Hispanic

white neighborhood comparison group. J Gerontol A Biol Sci Med Sci, 2000. 55(2): p. M90-7.

23. Tucker, K.L., O.I. Bermudez, and C. Castaneda, Type 2 diabetes is prevalent and poorly

controlled among Hispanic elders of Caribbean origin. Am J Public Health, 2000. 90(8): p. 1288-93.

24. Ranjit, N., et al., Socioeconomic Position, Race/Ethnicity, and Inflammation in the Multi-

Ethnic Study of Atherosclerosis. Circulation, 2007. 116(21): p. 2383-2390. 25. Renzaho, A.M., J.A. Halliday, and C. Nowson, Vitamin D, obesity, and obesity-related

chronic disease among ethnic minorities: a systematic review. Nutrition, 2011. 27(9): p. 868-79.



27. Bermudez, O.I., E.K. Becker, and K.L. Tucker, Development of sex-specific equations for

estimating stature of frail elderly Hispanics living in the northeastern United States. American Journal of Clinical Nutrition, 1999. 69(5): p. 992-8.

28. Chumlea, W., et al., Stature prediction equations for elderly non-Hispanic white, non-

Hispanic black, and Mexican-American persons developed from NHANES III data. J Am Diet Assoc, 1998. 98: p. 137 - 142.

29. Tucker, K., et al., Adaptation of a food frequency questionnaire to assess diets of Puerto

Rican and non-Hispanic adults. American Journal of Epidemiology, 1998. 148(5): p. 507 - 518.

30. Bermudez, O., et al., Hispanic and non-Hispanic white elders from Massachusetts have

different patterns of carotenoid intake and plasma concentrations. Journal of Nutrition, 2005. 135(6): p. 1496 - 1502.

31. Kwan, L.L., O.I. Bermudez, and K.L. Tucker, Low vitamin B-12 intake and status are

more prevalent in Hispanic older adults of Caribbean origin than in neighborhood-

matched non-Hispanic whites. Journal of Nutrition, 2002. 132(7): p. 2059-64. 32. Dawson-Hughes, B., et al., Estimates of optimal vitamin D status. Osteoporosis

International, 2005. 16(7): p. 713-716. 33. Holick, M.F., Vitamin D deficiency. N Engl J Med, 2007. 357(3): p. 266-81. 34. Document : PILKCR-7, 2003-11-25, High sensitivity CRP, and according to Immulate

1000 Operating Manual, Document #600467, rev B1, revision 5.XX August 2003, , D.P.C. (DCP), Editor 2003.

107

35. Jamal-Allial, A.A. and K.L. Tucker, Serum Vitamin D Concentrations, Dietary Sources,

and its relation to BMI among Puerto Ricans Living in the Greater Boston Area. To be submitted, 2013.

36. Giovannini, L., et al., 1,25-dihydroxyvitamin D3 dose-dependently inhibits LPS-induced

cytokines production in PBMC modulating intracellular calcium. Transplantation Proceedings, 2001. 33(3): p. 2366-2368.

37. Cohen, M., et al., Regulation of TNF-[agr] by 1[agr],25-dihydroxyvitamin D3 in human

macrophages from CAPD patients. Kidney International, 2001. 59(1): p. 69–75. 38. Alvarez-Rodriguez, L., et al., Age and low levels of circulating vitamin D are associated

with impaired innate immune function. Journal of Leukocyte Biology, 2012. 91(5): p. 829-838.

39. Nonn, L., et al., Inhibition of p38 by Vitamin D Reduces Interleukin-6 Production in

Normal Prostate Cells via Mitogen-Activated Protein Kinase Phosphatase 5:

Implications for Prostate Cancer Prevention by Vitamin D. Cancer Research, 2006. 66(8): p. 4516-4524.

40. Zittermann, A., et al., Vitamin D supplementation enhances the beneficial effects of

weight loss on cardiovascular disease risk markers. The American Journal of Clinical Nutrition, 2009. 89(5): p. 1321-1327.

41. Amer, M. and R. Qayyum, Relation between serum 25-hydroxyvitamin D and C-reactive

protein in asymptomatic adults (from the continuous National Health and Nutrition

Examination Survey 2001 to 2006). American Journal of Cardiology, 2012. 109(2): p. 226-30.

42. Falcon, L.M. and K.L. Tucker, Prevalence and correlates of depressive symptoms among

Hispanic elders in Massachusetts. J Gerontol B Psychol Sci Soc Sci, 2000. 55(2): p. S108-16.

43. Fantuzzi, G., Adipose tissue, adipokines, and inflammation. The Journal of allergy and clinical immunology, 2005. 115(5): p. 911-9; quiz 920.

44. Van Lente, F., Markers of inflammation as predictors in cardiovascular disease. Clin Chim Acta, 2000. 293(1-2): p. 31-52.

45. Dietary Reference Intakes for Calcium and Vitamin D. Committee to Review Dietary

Reference Intakes for Vitamin D and Calcium. , in 2010, Institute of Medicine: Washington, DC: National Academy of Sciences.

108

TABLE 2.1 Descriptive characteristic of the men and women of The Boston Puerto Rican

Osteoporosis Study

Variables Women (n=699 )

Men (n=273 )

P-value

Age (y) 60.3 ± 7.5 59.5 ± 8.0 0.2

Education (above 8th grade %) 49.0 44.3 0.1

Poverty (%) 59.0 49.2 0.009

Serum 25(OH)D (nmol/L) 50.1 ± 19.0 47.0 ± 17.5 0.01

Adiposity Variables

BMI (kg/m2) 33.0 ± 7.0 30.2 ± 5.5 <0.0001

Waist Circumference (cm) 104 ± 15 104 ± 14 0.4

Total Fat Mass (kg) 36.0 ± 1.5 26.5 ± 1.1 <0.0001

Android Fat mass (kg) 3.4 ± 1.3 3.0 ± 1.4 0.0007

Inflammatory Markers

Serum hsCRP (mg/L) $ 3.5± 3.3 2.3 ± 3.2 <0.0001

Serum TNF-α (pg/ml) $ 3.0 ± 1.5 3.1 ± 1.5 0.6

Serum IL-6 (pg/ml) $ 3.2 ± 2.3 3.1 ± 2.1 0.9

WBC x103/mm3 6.7 ± 2.1 6.8 ± 2.0 0.8

Other Covariates

Use of NSAIDs (%) 34.0 18.0 <0.0001

Current- smoking (%) 18.0 30.0 <0.0001

109

Current- alcohol drinking (%) 28.2 42.5 <0.0001

Physical Activity Score$‡ 31.3 ± 4.0 33.0 ± 6.0 <0.0001

Miles walking/day$‡ 1.4 ± 2.0 2.0 ± 2.4 0.0002

Dietary Measures

Vitamin D Supplement (>200IU/day) # (%)

30.2 18.5 0.0002

Total vitamin D intake (IU/day) || 288 ± 152 257 ± 123 0.002

Total energy intake (kcal/day) 1770 ± 825 2250± 1200 <0.0001

BMI, body mass index; 25(OH)D, 25-hydroxyvitamin D. 1Data are presented as mean ± SD for continuous variables or as % for categorical variables $Geometric mean ± SD Means were compared using t-tests. Percentages were compared using the Chi-square test (χ2) . ‡ The range for the physical activity score = 0-35.5 for Puerto Rican men and 26.4 for Puerto Rican women. Range for Miles/day =0- 22 for men and 0-20 miles/day for Puerto Rican women. #The recommended dietary allowance (RDA) for vitamin D for men and women aged 19-70 years =15 ug /day [82] ||Adjusted for energy intake using ANOVA (PROC GLM; SAS Institute, Cary, NC).

110

TABLE 2.2 Relationships between serum 25(OH)D concentration and inflammatory markers

Inflammatory Markers Serum 25(OH)D concentration

r p-value

Serum hsCRP (mg/L) - 0.05 0.1

Serum TNF-α (pg/ml) - 0.1 0.0005

Serum IL-6 (pg/ml) - 0.11 0.001

Pearson partial linear correlation coefficient (r) is age and sex adjusted 25(OH)D, 25-hydroxyvitamin D; hsCRP, high-sensitivity C-reactive protein; TNF-α, Tumor necrosis factor-alpha; IL-6, Interleukin 6.

111

TABLE 2.3: Adjusted associations between inflammatory markers and 25(OH)D concentration.

Serum 25(OH)D (nmol/L)

IL-6 TNF-α

β 1 SE p β SE p

Risk Factor - 0.003 0.001 0.01 - 0.002 0.001 0.02

Obesity - 0.003 0.001 0.01 - 0.0015 0.001 0.03

Life Style - 0.0024 0.001 0.03 - 0.002 0.001 0.02

1Regression coefficients (± SE) Risk Factor Model: Sex, age seasonality, WBC count (mm3), and use of Aspirin (Y/N) Obesity Model: Risk factor model + BMI, total fat-mass and android fat-mass Life Style Model: Risk-factor-model+ Obesity-model, smoking and Physical Activity Score

112

TABLE 2.4: Regression coefficients (± SE) from multivariate regression analysis of hsCRP

(mg/dL) and serum 25(OH)D concentration.


hsCRP (mg/dL)

β SE p

Risk Factor Model 0.002 0.002 0.3

Obesity Model 0.001 0.001 0.3

Comprehensive Model 0.0005 0.001 0.5

Risk Factor Model: Sex, age seasonality, WBC count (mm3), and use of non-steroidal anti-inflammatory (Y/N), energy intake (kcal/day) smoking (Y/N). Obesity Model: Risk factor model + BMI, total fat-mass (g) and android fat-mass (g), diastolic blood pressure and diabetes (Y/N) Comprehensive Model: Risk-factor-model+ Obesity-model, poverty, and education.

113

FIGURE 2.1 Adjusted means (±SEM) of IL-6 concentration (pg/ml) by tertiles of adjusted

serum 25(OH)D concentration among 893 Puerto Rican men and women. The LS-Means

adjusted for sex, age, BMI, total fat-mass, android fat-mass, physical activity, and smoking.

* Different from referent group (25(OH)D < 40 nmol/L) (Dunnett adjustment for multiple

comparisons; P < 0.05)

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5

< 40 nmol/L 41-55 nmol/L > 56 nmol/L

Seu

m I

L-6

con

cen

trati

on

(p

g/m

l)

Vitamin D Tertiles (nmol/L)

P-trend = 0.002

**

114

FIGURE 2.2 Adjusted means (±SEM) of TNF-α concentration (pg/ml) by tertiles of adjusted


adjusted for sex, age, BMI, total fat-mass, android fat-mass, physical activity, and smoking.

* Different from referent group (25(OH)D < 40 nmol/L) (Dunnett adjustment for multiple

comparisons; P < 0.05)

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5


Seu

m T

NF

-αco

nce

ntr

ati

on

(p

g/m

l)


P-trend = 0.001

*

115

FIGURE 2.3 Adjusted means (±SEM) of hsCRP concentration (mg/dL) by tertiles of adjusted


adjusted for sex, age, BMI, total fat-mass, WC count, use of NSAID’s, physical activity, and

smoking. Dunnett adjustment for multiple comparisons; P > 0.05

0

0.5

1

1.5

2

2.5

3

3.5

4

4.5

5


Ser

um

hs-

CR

P

con

cen

trati

on

(m

g/d

L)


116

Poster Presentation:

PDF of the presentation of Chapter Two data at Experimental Biology 2013

117

CHAPTER FOUR

Longitudinal Association of Vitamin D Serum Concentration & Adiposity

Phenotypes


1Department of Pharmaceutical Science. 3Department of Health Sciences, 2Bouvé College of

Health Sciences. Northeastern University.


Submitted to the Journal of Steroid Biochemistry & Molecular Biology

118

ABSTRACT

Vitamin D deficiency is prevalent, especially among ethnic minorities. There are several

cross-sectional studies on the association between serum 25-hydroxy vitamin D concentration

(25(OH)D) concentration and adiposity measures such as body mass index (BMI), and waist

circumference (WC). We examined the longitudinal association between BMI and WC on serum

25(OH)D concentration in 970 Puerto Rican adults: 270 men and 700 women, aged 45-75 years.

Our analyses showed negative linear correlations at two time points between BMI (r = -0.1, P

=0.02) and WC (r = -0.1, P =0.009) and serum 25(OH)D concentration. Multivariate analysis

showed that baseline-BMI was significantly inversely associated with change in serum 25(OH)D

concentration (P<0.02), controlling for age and sex. This association remained significant after

adjusting for vitamin D supplement use, smoking, physical activity and alcohol intake. Baseline-

WC was not significantly associated with change in serum 25(OH)D concentration. The major

finding of the present study is that obesity appears to have a negative consequence on serum

25(OH)D concentration in this population of older Puerto Rican adults living in the Boston area,

which might accelerate vitamin D deficiency status.

119

ABBREVIATIONS:

25-hydroxyvitamin D (25(OH)D); cardiovascular disease (CVD); confidence interval (CI);

United States (USA); food-frequency questionnaire (FFQ); body mass index (BMI)

120

INTRODUCTION

The prevalence of overweight and obesity has increased among all socioeconomic

classes, leading to an increased risk of associated mortality and morbidity [167]. A recent report

suggested that vitamin D deficiency is another re-emerging global health problem that might

further contribute to an increase in the prevalence of related-chronic diseases [168]. Several

factors are known to influence 25-hydroxy vitamin D [25(OH)D] concentration, including age,

sex, sun exposure, skin pigmentation, ethnicity, adiposity, and physical activity [55, 56].

Ethnic minorities, generally, and Puerto Ricans, specifically, shoulder the greatest burden

of health disparities, including a greater risk of chronic disease [109-111]. In the United States

obesity continues to rise, specifically among minorities such as Hispanics, i.e. 39.1% [169, 170].

Additionally, a systematic review concluded that ethnic minorities have a significantly higher

risk of vitamin D deficiency among all age groups. They further concluded that there is a number

of indications that link vitamin D deficiency to obesity-related chronic disease [112].

The association between vitamin D deficiency and obesity has been consistently

documented, indicating that obesity is inversely associated with serum 25(OH)D [132, 171-173].

One of the postulated theories to explain this phenomenon is the sequestering of vitamin D into

adipose tissue, thus affecting its bioavailability for further metabolism [174, 175]. Another is that

obese individuals tend to have lower outdoor physical activity and greater time spent indoors

[173]. In the cross-sectional Tromsø study, BMI was significantly inversely associated with

serum 25(OH)D [132]. The Insulin Resistance Atherosclerosis (IRAS) Family study reported

that among Hispanics, 25(OH)D was inversely associated with adiposity phenotype; however, it

was not associated with 5-yr change in adiposity [176]. Another study demonstrated that serum

121

25(OH)D concentration was inversely associated with BMI (r = -0.18, P =0.005) and WC (r = -

0.14, P =0.03) [177].

The aim of this study was to determine the longitudinal association between serum

25(OH)D concentration in relation to BMI and WC among participants of the Boston Puerto

Rican Health Study (BPRHS) and the Boston Puerto Rican Osteoporosis Study (BPROS) in the

Greater Boston Area.



Recruitment and methods of the BPRHS [83] and its ancillary study, the BPROS [113],

are described in detail elsewhere. Concisely, at both time points, a bilingual interviewer

conducted home visits and administrated detailed questionnaires on general background,

socioeconomic status, education, household income, health and health behaviors that include

medical diagnosis, detailed use of medication(s), depression and a psychological acculturation

scale. They also answered questionnaires on dietary intake, food security, activities of daily

living, migration status, stress, depressive symptomology, and cognitive function. Moreover, the

interviewer collected information about anthropometrics, physical performance measures and

blood pressure. The following day, the study phlebotomist collected 12-hr urine, 12-hr fasting

blood and saliva. By December 2010, 1504 participants completed baseline visits. Once the 2-yr

follow up was completed, participants were re-consented for the BPROS. By May 2013, 1257

participants completed 2year follow-up visits and 970 of which completed the BPROS study.

These participants contributed 24-hr urine measures as well. All study protocols were approved

122

by the Institutional Review Boards at Tufts Medical Center and Northeastern University, and

each participant provided written informed consent.

There were no significant differences in socio-demographic characteristics between those

with complete or incomplete information. The present study includes complete longitudinal data

for 700 Puerto Rican women and 270 Puerto Rican men.


Enrolled participants completed questionnaires that captured information on general

background and socioeconomic status, dietary intake, food security, health history, health

behaviors, health insurance, medical conditions and detail use of medication(s), activities of

daily living instrument, migration status, stress, depressive symptomology, and cognitive

function. Moreover, anthropometric and blood pressure measures, and physical function tests,

were completed. Weight, and waist and hip circumferences were measured in duplicate. Weight

was measured using a clinical scale (Toledo Weight Plate, Model I5S, Bay State, and Systems

Inc. Burlington, MA). Body mass index (BMI) was calculated as weight (kg) divided by height

(m) squared [27, 28]; BMI was categorized into: recommended (BMI <25), overweight (25-30),

and obese (≥ 30) [43]. For the purpose of this analysis, vitamin D deficiency was defined as

serum 25(OH)D concentration below 50 nmol/L, insufficiency between 50-75 nmol/L, and

sufficiency above 75 nmol/L [44, 45]. Seasonality was defined as summer, July to September;

fall, October to December; winter, January to March; and spring, April to June. Smoking status

was categorized into yes and no for the current behavior. Vitamin D from supplements was

defined as: minimal 0-10 µg/day (200-400 IU/day); and high ≥ 10 µg/day (≥ 400 IU/day).

123

Education level was measured at baseline and defined using the cutoff point of 8th grade while

poverty was measured at each time point and defined following the U.S. Department of Health

and Human Services. Dietary vitamin D intake was estimated from a semi-quantitative food-

frequency questionnaire (FFQ) that was specifically designed for this population [46]. This FFQ

has been validated against serum carotenoids [47], and vitamin B12 [48] in a Hispanic

population aged ≥ 60 years. Those with energy intakes < 600 or > 4800 kilocalories and/or > 10

questions blank on the FFQ were excluded from the analyses. Nutrient intakes were calculated

with the Nutrition Data System for Research (NDS-R) software version 2007 (Nutrition

Coordinating Center, University of Minnesota, Minneapolis, MN, Food and Nutrient Database

2007, released 2007). Alcohol intake (gm) was captured by the FFQ. Heart disease status was

self-reported by the participants.

Laboratory assays



saved and stored at -80°C until processed. Serum 25(OH)D concentration was measured by

extraction followed by 25I radioimmunoassay Packard COBRA II Gamma Counter (DiaSorin

Inc., Stillwater, MN 55082) catalog # 68100E with intra-assay and inter-assay CV% of 10.8%

and 9.4%, respectively. Serum creatinine was assessed with a colorimetric, kinetic reaction using

the Olympus AU400e with Olympus Creatinine Reagents (OSR6178) (Olympus America Inc.,

Melville, NY) with intra-assay and inter-assay CV% of 2.0%, and 4.0%, respectively.

124


All statistical analyses were conducted with SAS 9.1.3 (SAS Institute Inc., Cary, NC);

and all tests were 2-sided with P <0.05 considered statistically significant. Dependent skewed

variables were natural log-transformed to normalize their distributions before the analysis.

Demographic and clinical characteristics were compared using student t-test for continuous

variables and chi square for categorical variables. Characteristics were compared by sex using

two-tailed t-test for two independent samples, and with χ2 test for differences in proportion.

Paired t-test was used to compare characteristics across time points. Pearson partial linear

correlation analysis was used to examine the relationship between serum 25(OH)D concentration

with adiposity phenotype at each time point, adjusting for age and sex. We constructed general

linear models (PROC GLM) to model longitudinal associations between the change in 25(OH)D

concentration (continuous) and adiposity phenotype (continuous). Because adiposity phenotype

might be sex-specific, we constructed sex-stratified analyses in addition to the full model, even if

the by-sex interaction term was not significant. Model-1 was adjusted for age (y), baseline-

25(OH)D (nmol/L/), serum creatinine (µmol/L), follow-up time (months), and sex in the full

model. In addition to model-1 covariates, model-2 included seasonality, dietary vitamin D &

supplements intake. Model-3 was adjusted for model-1 and 2 covariates plus physical activity,

alcohol intake (g), current smoking status (y/n), poverty level and education. Least-square means

and linear trends of 25(OH)D across the BMI categories were assessed using PROC GLM and

adjusted for multiple comparisons using Dunnett’s adjustment, with the lowest group as the

reference group. The model was repeated for each time point, independent of the other. The

model was adjusted for age (y), sex, seasonality, current smoking status (y/n), total energy intake

(kcal/d), physical activity, serum creatinine, dietary vitamin D & vitamin D supplement intake.

125

For all multivariate regression models, we checked the assumptions of linearity and

homogeneity through examining the outcome residuals against the exposure. For all final

models, we performed influence diagnostics tests to check for outliers and influential points. We

also tested all of our models for potential effect modifications by sex, with covariates such as

smoking, by including interaction terms in the model. The baseline concentration of each lipid

profile component was included as a covariate to all of our models to control for regression to

the mean [121]. To reduce bias and augment precision, we conducted multiple imputation

analysis to consider the uncertainty that missing values might convoy within our dataset. We

created a total of five imputed data sets with (seed= 32, 37854) using procedure (PROC MI); we

included all covariates that are considered risk factors and being used in any of our models. We

used (PROC MIANALYZE) to combine all the five imputed data sets, thus, producing valid

statistical inferences that accounts for within and between imputation variation [122]. We found

that these missing values produced no significant impact on our analysis.

RESULTS

Serum 25(OH)D concentration was available for 650 women and 255 men at baseline

and 2-yr. The mean ages of men and women were 59.5 ± 8.0 and 60.5 ± 8.0, respectively. The

mean follow-up time for women was 33.3± 12 months while men had a mean follow-up time

equal to 30.5 ± 49 months. At baseline, the mean serum 25(OH)D of the women and men was,

44.0 ± 17.5 and 43.0 ± 16.1 nmol/L, respectively, while the 2-year 50.1 ± 19.0 and 47.0 ± 17.5

nmol/L, respectively. Among all participants, the mean of 25(OH)D at 2-yr was 5.5 nmol/L

higher compared to baseline (P<0.0001). Compared to men, women at the 2-yr time point had

126

significantly higher serum 25(OH)D (p=0.02). Figure 3.1 shows the monthly distribution of

serum 25(OH)D for the two time points.

While BMI values did not differ between the two time points within each sex, Puerto

Rican women did exhibit significantly higher BMI compared to Puerto Rican men, i.e. 33.0 and

30 (P < 0.0001), respectively. On the contrary, WC did not differ between both sex, but it was

significantly higher at 2-yr compared to baseline (P < 0.0001). No significant change in the

participants’ dietary vitamin D intake. However, Puerto Rican women had higher dietary vitamin

D intake (IU/day) and higher use-rate of VD-containing supplements. Overall men had higher

energy intake (kcal/day) compared to women at both time points. On the other hand, there was a

clear significant reduction in total calories consumed, i.e. µ=365 kcal/day (P < 0.001) within the

2.5 years of follow-up.

Table 2 shows the age and sex Pearson linear correlation coefficients of serum 25(OH)D

concentration and adiposity phenotype. BMI and WC showed significant negative linear

correlation with serum 25(OH)D serum 25(OH)D at both time points. Baseline-BMI (r = -0.1, P

=0.006) and Baseline-WC (r = -0.1, P =0.02), 2yr-BMI (r = -0.1, P =0.01), 2yr-WC (r = -0.1, P

=0.01).

The longitudinal linear regression model (Table 3) showed that baseline-BMI was

significantly inversely associated with change in serum 25(OH)D concentration (β = -0.17, P

=0.04) controlling for age, sex, creatinine, and follow-up time. In model-2, the inclusion of

seasonality, dietary vitamin D & supplements did not alter the regression coefficient or its

significance. Nevertheless, the additional adjustment of the association for miles/day, alcohol,

smoking, poverty index, and education status (model-3), strengthened the inverse association (β

= - 0.21, P =0.01). The adjusted-mean serum 25(OH)D was significantly lower with increased

127

obesity among the participants at both time points, P-trend <0.01 and <0.002, respectively (Fig.

1). At baseline, participants with BMI ≥ 40 had significantly lower 25(OH)D (38.5 ± 1.5 nmol/L)

compared to BMI ≤ 25 (44.5 ± 1.2 nmol/L), while at the 2-yr time point, participants with BMI

who were classified as obese had significantly lower 25(OH)D (45.5 ± 1.2 nmol/L) compared to

BMI ≤ 25 (50.1 ± 1.5 nmol/L).

DISSCUSION

The results of our longitudinal analysis confirm the inverse association between low

serum 25(OH)D concentration and BMI. Our observation of lower adjusted mean serum

25(OH)D concentration in obese, severely obese and morbidly obese participants is in agreement

with many studies [132, 172, 176, 177]. Our longitudinal multivariate linear regression showed

that baseline-BMI was negatively associated with the change in serum 25(OH)D concentration

(P =0.01), independent of covariates including age, sex, serum creatinine, physical activity,

smoking, alcohol, seasonality, poverty and education. Conversely, we did not observe similar

longitudinal associations with WC, although it was significantly correlated with serum 25(OH)D

at both time points.

The mechanism that can explain the phenomena of low serum 25(OH)D and increased

adiposity risk is not clear. A review by Earthman and colleague discusses a number of proposed

theories that link Vitamin D deficiency to adiposity [178]. One theory suggests that because

Vitamin D is fat-soluble, it might be sequestered within adipose tissue. They further suggest that

the adipose tissue of obese individuals might have higher 25(OH)D uptake and storage relative to

lean individuals [48, 178]. Another theory suggests that due to the fact that adiposity, especially

128

visceral adiposity, contributes to a high inflammatory state, this in turn, contributes to lower

serum 25(OH)D concentration [178, 179]. Another hypothesis suggests that because of increased

adiposity, there may possibly be an increase of the action of 24-hydroxylase, leading to amplified

catabolism of vitamin D [178, 180]. Consequently, further research is needed to elucidate our

understanding of how vitamin D may be affected by adiposity or how vitamin D affects obesity.

More than 60% of the participants of BPROS had serum 25(OH)D concentration below

50 nmol/L, which is considered deficient, according to the Endocrine Society [115]. A recent

report by the National Health and Nutrition Examination Survey (NHANES) III, showed that

more than 42% of their 2001-2006 participants had a vitamin D concentration below 50 nmol/L;

when stratified by ethnicity, Hispanics were second highest in vitamin D deficiency with 70%

[181]. Our prior report indicated that among 1502 participants of the BPRHS, more than 70%

had serum 25(OH)D concentration below 50 nmol/L at baseline, concurring with the NHANES

III data [159]. This greater risk of vitamin D deficiency may be related to greater risk of health

disparities [182]. A report by Palacios and colleagues showed that Puerto Ricans who are

residing in Puerto Rico, had higher serum 25(OH)D concentration, approximately 75 nmol/L,

compared to our population, at baseline 43.4 ± 17.2 nmol/L [159] and 2yr 50.0 ± 19.0 nmol/L.

This difference in serum 25(OH)D concentration is likely due to sun exposure, generally, and

latitude, specifically, i.e. Puerto Rico is at 18.26° while Boston is at 42.35° [26].

The present study has limitations: we did not have direct measures of sun exposure at

baseline. However, our study has substantial strengths: a total large sample size with low attrition

and the use of a detailed questionnaire with an FFQ that allowed detailed information.

129

In summary, serum 25(OH)D concentration was inversely associated with 2.5-yr change

in BMI. The adjusted-mean serum 25(OH)D concentration showed a declining trend, at both

time points, with greater BMI, especially when reaching class III obesity.

130

BIBLIOGRAPHY

1. Organization, W.H., Noncommunicable diseases country profiles 2011. WHO global

report, 2011. p. 209.

2. Mithal, A., et al., Global vitamin D status and determinants of hypovitaminosis D.

Osteoporosis International, 2009. 20(11): p. 1807-20.

3. Holick, M.F., Vitamin D status: measurement, interpretation, and clinical application.

Ann Epidemiol, 2009. 19(2): p. 73-8.

4. Holick, M.F. and T.C. Chen, Vitamin D deficiency: a worldwide problem with health

consequences. Am J Clin Nutr, 2008. 87(4): p. 1080S-6S.


Puerto Rican Health Study. Social Science & Medicine, 2010. 70(12): p. 1988-96.


limitation among Massachusetts elderly Puerto Ricans, Dominicans, and non-Hispanic white

neighborhood comparison group. J Gerontol A Biol Sci Med Sci, 2000. 55(2): p. M90-7.


controlled among Hispanic elders of Caribbean origin. Am J Public Health, 2000. 90(8): p.

1288-93.

8. Ogden CL, C.M., Kit BK, Flegal KM., Prevalence of obesity in the United States, 2009–

2010. NCHS data brief, 2012, National Center for Health.

9. Flegal Km, C.M.D.K.B.K.O.C.L., Prevalence of obesity and trends in the distribution of

body mass index among us adults, 1999-2010. JAMA, 2012. 307(5): p. 491-497.

10. Renzaho, A.M., J.A. Halliday, and C. Nowson, Vitamin D, obesity, and obesity-related


11. Renzaho, A.M.N., J.A. Halliday, and C. Nowson, Vitamin D, obesity, and obesity-related

chronic disease among ethnic minorities: A systematic review. Nutrition, 2011. 27(9): p. 868-

879.


hydroxyvitamin D and body mass index: the Tromsø study. European Journal of Nutrition, 2010.

49(7): p. 401-407.

13. Snijder, M.B., et al., Adiposity in relation to vitamin D status and parathyroid hormone

levels: a population-based study in older men and women. Journal of Clinical Endocrinology &

Metabolism, 2005. 90(7): p. 4119-23.

14. Brock, K., et al., Low vitamin D status is associated with physical inactivity, obesity and

low vitamin D intake in a large US sample of healthy middle-aged men and women. J Steroid

Biochem Mol Biol, 2010. 121(1-2): p. 462-6.

15. Norman, A.W., Minireview: vitamin D receptor: new assignments for an already busy

receptor. Endocrinology, 2006. 147(12): p. 5542-8.

131

16. Tuohimaa, P., Vitamin D and aging. Journal of Steroid Biochemistry & Molecular

Biology, 2009. 114(1-2): p. 78-84.

17. Young, K.A., et al., Association of Plasma Vitamin D Levels with Adiposity in Hispanic

and African Americans. Journal of Clinical Endocrinology & Metabolism, 2009. 94(9): p. 3306-

3313.

18. McGill, A.-T., et al., Relationships of low serum vitamin D3 with anthropometry and

markers of the metabolic syndrome and diabetes in overweight and obesity. Nutrition Journal,

2008. 7: p. 4.

19. Tucker, K., et al., The Boston Puerto Rican Health Study, a longitudinal cohort study on

health disparities in Puerto Rican adults: challenges and opportunities. BMC Public Health,

2010. 10(1): p. 107.


mineral density in older Puerto Rican adults. The American Journal of Clinical Nutrition, 2011.

94(4): p. 1063-1070.

21. Holick, M.F., et al., Evaluation, Treatment, and Prevention of Vitamin D Deficiency: an

Endocrine Society Clinical Practice Guideline. Journal of Clinical Endocrinology &

Metabolism, 2011. 96(7): p. 1911-1930.

22. Dawson-Hughes, B., et al., Estimates of optimal vitamin D status. Osteoporos Int, 2005.

16(7): p. 713-6.

23. Flegal, K.M., et al., Overweight and obesity in the United States: prevalence and trends,

1960-1994. International journal of obesity and related metabolic disorders : journal of the

International Association for the Study of Obesity, 1998. 22(1): p. 39-47.




different patterns of carotenoid intake and plasma concentrations. Journal of Nutrition, 2005.

135(6): p. 1496 - 1502.


more prevalent in Hispanic older adults of Caribbean origin than in neighborhood-matched non-

Hispanic whites. Journal of Nutrition, 2002. 132(7): p. 2059-64.

27. Earthman, C.P., et al., The link between obesity and low circulating 25-hydroxyvitamin D

concentrations: considerations and implications. Int J Obes, 2012. 36(3): p. 387- 96.

28. Wortsman, J., et al., Decreased bioavailability of vitamin D in obesity. The American

Journal of Clinical Nutrition, 2000. 72(3): p. 690-693.

29. Compher, C. and K.O. Badellino, Obesity and Inflammation: Lessons From Bariatric

Surgery. Journal of Parenteral and Enteral Nutrition, 2008. 32(6): p. 645-647.

30. Li, J., et al., 1α,25-Dihydroxyvitamin D hydroxylase in adipocytes. The Journal of Steroid

Biochemistry and Molecular Biology, 2008. 112(1–3): p. 122-126.

132

31. Looker, A.C.J., C L. Lacher, D A. Pfeiffer, C M. Schleicher, R L. Sempos, C T.,

Vitamin D status: United States, 2001–2006, in NCHS data brief2011, National Center for

Health Statistics: Hyattsville, MD.


and its relation to BMI among Puerto Ricans Living in the Greater Boston Area. To be

submitted, 2013.

33. Forrest, K.Y.Z. and W.L. Stuhldreher, Prevalence and correlates of vitamin D deficiency

in US adults. Nutrition Research, 2011. 31(1): p. 48-54.


133

TABLE 3.1 Descriptive characteristic of men and women in the Boston Puerto Rican Health

Study (BPRHS) & Boston Puerto Rican Osteoporosis Study (BPROS)

Baseline (n=1504)

Two Year (n=970)

Women (n=1050)

Men (n=445)

Women (n=700 )

Men (n=270)

Age (y) 57.0 ± 7.3 56.0 ± 8.0 60.3 ± 7.5 59.5 ± 8.0

Education (>8th grade %) 51.5 55.5 -- --

Poverty (%) 60.0 49.0$$ 59.0 48.5$$

Serum 25(OH)D (nmol/L) §§§ 44.0± 17.4 43.0 ± 16.5 50.1 ± 19.0 47.0 ± 17.5*

Follow-up time (months) -- -- 33.3± 12 30.5± 49

BMI (kg/m2) 33.2 ± 7.0 30.0 ± 5.3*** 33.0 ± 7.0 30.2 ± 5.5***

Waist Circumference (cm) §§§ 102.2 ± 15.4 103 ± 14 104 ± 15 104 ± 14

Miles walking/day§§§ 1.0 ± 1.5 1.4 ± 2.0*** 1.4 ± 2.0 2.0 ± 2.4 ***

Current- smoking (%) 20.1 32.5$$$ -- --

Current- alcohol drinking (%) 36.0 51.0$$$ -- --

Vitamin D Supplement (>200IU/day) (%)

30.0 24.1 30.2 18.5$$

Total vitamin D intake (IU/day) 291 ± 172 280 ± 178 288 ± 152 257 ± 123.*

BMI, body mass index; 25(OH)D, 25-hydroxyvitamin D. 1Data are presented as mean ± SD for continuous variables or as % for categorical variables Means were compared using t-test for between sex (within each time points): *P -value <0.05, ** P -value <0.001, ***P-value <0.0001

Means were compared using paired t-test for between time points: §P -value <0.05,

§§P -value

<0.001, §§§

P-value <0.0001

Percentages were compared using Chi-square test for between sex (within each time points): $P -value <0.05, $$

P -value <0.001, $$$ P-value <0.0001

134

TABLE 3.2 Pearson partial linear correlation of serum 25(OH)D concentration and adiposity

phenotype at baseline and 2-year

Baseline Two-Year

25(OH)D p-value 25(OH)D p-value

BMI -0.1 0.006 -0.1 0.01

WC -0.1 0.02 -0.1 0.01

Pearson partial linear correlation coefficient (r) is age and sex adjusted Abbreviations: BMI, body mass index; 25(OH)D, 25-hydroxyvitamin D; WC, waist circumference; AFM, abdominal fat mass; TBF, total body fat; %TBF, total body fat percentage

135

TABLE 3.3 Regression coefficients (± SE) of adiposity phenotypes on plasma 25(OH)D concentration

Adiposity Phenotype

Change in 25(OH)D (nmol/L)

Baseline-BMI β ± SE P-value

Baseline-WC β ± SE P-value

ALL (n=866)

Men (n=246)

Women (n=620)

ALL

(n=864) Men

(n=245) Women (n=619)


Model 1 -0.17 (± 0.1)

0.04 -0.25 (± 0.2)

0.1 -0.13 (± 0.1)

0.1 -0.03 (± 0.03)

0.3 -0.06 (± 0.07)

0.3 -0.02 (± 0.04)

0.5

Model 2 -0.22 (± 0.1)

0.01 -0.26 (± 0.2)

0.1 -0.15 (± 0.1)

0.1 -0.03 (± 0.03)

0.3 -0.06 (± 0.1)

0.3 -0.02 (± 0.04)

0.6

Model 3 -0.22 (± 0.1)

0.01 -0.3 (± 0.2)

0.1 -0.20 (± 0.1)

0.04 -0.03 (± 0.03)

0.3 -0.06 (± 0.1)

0.3 -0.03 (± 0.04)

0.4

Model 1: adjusted for age, baseline 25(OH)D, serum creatinine, time f/up (mos), and sex in the case of the combined data1 Model 2: adjusted for model 1 variables plus seasonality, dietary vitamin D & supplements and sex in the case of the combined data1

Model 3: adjusted for model 1 variables plus physical activity, alcohol, smoking, poverty level, education and sex in the case of the combined data1

136

FIGURE 3.1 The monthly distribution of serum 25(OH)D for the participants of BPRHS at both time points

137

FIGURE 3.2 The adjusted means of serum 25(OH)D concentration within the categories of class III obesity LS-Means adjusted for Sex, age seasonality, miles/day, vitamin D supplements, dietary VD, smoking, and total calories. * Different from referent group (BMI ≥ 25) (Dunnett adjustment for multiple comparisons; P < 0.05)

35

37

39

41

43

45

47

49

51

53

55

25 & lessIdeal

25-30Overwieght

30-35 Obese 35-40 Severeobesity

40 & moreMorbidobesity

Ser

um

25(O

H)D

nm

ol/

L

BMI Categories

Baseline

Two-Year

*

*

P-trend < 0.002

P-trend < 0.01

138

Poster Presentation:

PDF of the presentation of Chapter Three data at Vitamin D Workshop 2013

139

CHAPTER FIVE

The Association of Total and Regional Adiposity with Serum 25-

hydroxyvitamin D Concentration

The Boston Puerto Rican Osteoporosis Study (BPROS)


1Department of Pharmaceutical Science Department. 3Department of Health Sciences,

2Bouvé College of Health Sciences. Northeastern University.


140

ABSTRACT

Vitamin D deficiency is prevalent, especially among ethnic minorities. There are

several cross-sectional studies on the association between serum 25-hydroxyvitamin D

concentration (25(OH)D) and adiposity phenotype; specifically: total body fat (TBF),

percent body fat (%TBF), android fat-mass (AFM) and gynoid fat-mass (GFM), using

dual-energy x-ray absorptiometry (DXA). We examined this association among 970

Puerto Rican adults, 270 men and 700 women; aged 45-75, years from the Boston Puerto

Rican Osteoporosis study. Our analyses showed that TBF-in kilograms-or %TBF was

not significantly associated with serum 25(OH)D. Moreover, AFM showed inverse

significant association with serum 25(OH)D in both Puerto Rican men and women,

whereas GFM showed a positive significant association with serum 25(OH)D among

Puerto Rican women only. This suggests that obesity is negatively associated with serum

25(OH)D concentration in this population of older Puerto Rican adults living in the

Boston area.

141

ABBREVIATIONS:

25-hydroxyvitamin D concentration (25(OH)D); android fat-mass (AFM); cardiovascular

disease (CVD); 95% confidence interval (95% CI); dual-energy x-ray absorptiometry

(DXA);gynoid fat-mass (GFM) ;Life-Style model (LS-model); Obesity model (O-model);

Parathyroid Hormone, Intact (iPTH); percent body fat (%TBF); Risk Factor model (RF-

model); total body fat (TBF); United States (USA).

142

INTRODUCTION

Ethnic minorities, generally, and Puerto Ricans, specifically, are shown to carry a

greater burden of health disparities than non-Hispanic whites, which results in a greater

risk of chronic disease [109-111]. Obesity in the United States (USA) is on the rise,

especially among minorities such as Hispanics, i.e. 38.7% (95% confidence interval [CI],

33.5%- 43.9%) [169, 170, 183]. A 2010 report by Flegal indicated that Hispanic men and

women exhibit different prevalence of obesity after the age 40; 37.4% (95% CI, 29.0-

45.8) and 46.6% (95% CI, 37.3-55.9), respectively [183]. Casa and colleagues found that

ethnicity might be associated with moderately higher prevalence of obesity compared to

non-Hispanic whites; with about a 5% higher BMI. They also reported that percentage

body-fat was about 4% greater among Hispanic, relative to non-Hispanic white, women-

especially in the arms and trunks areas [184].

Android obesity, or central obesity, is when the predominant fat deposition is in

the abdomen, while gynoid obesity is when the predominant fat deposition occurs in the

hip area [185]. Android obesity is thought to be related to increased risk of cardiovascular

disease (CVD), while gynoid obesity is less related to CVD [186]. Whereas most of the

research on the association between serum 25(OH)D concentration and adiposity

measures have focused on BMI, it is not clear how AFM and GFM are associated with

serum 25(OH)D concentration.

Vitamin D deficiency is emerging as a global health problem that may lead to an

increase of chronic disease, including CVD, osteoporosis and other immune-related

diseases [56, 168]. Several factors are known to influence serum 25(OH)D concentration,

143

including age, sex, sun exposure, skin pigmentation, ethnicity, adiposity, and physical

activity [49, 55, 56]. It is suggested that some ethnic minorities are more susceptible to an

increased risk of vitamin D deficiency than others [176, 187]. In a systematic review,

Renzaho and colleagues discuss that ethnic minorities have a significantly higher risk of

vitamin D deficiency among all age groups. They conclude that there is a number of

indications linking vitamin D deficiency to obesity-related chronic disease among

minorities [112]. Our prior report showed that the mean serum 25(OH)D concentration

in Puerto Rican adults in the Greater Boston Area was 43.5 ± 16.5 nmol/L [159]. One

postulated theory to explain this phenomenon is the sequestering of vitamin D into

adipose tissue, thus affecting its bioavailability for further metabolism [49, 174, 175].

Another is that people who are obese tend to have lower outdoor physical activity, and

increased time spent indoors [173]. Numerous studies have investigated the association

between adiposity and serum 25(OH)D concentration, applying different indexes of

adiposity [48, 60, 64, 132, 172, 173, 176, 184, 188-190]. The Longitudinal Aging Study

Amsterdam showed that TBF was inversely associated with 25(OH)D [172]. A study in

south-east Texas showed that among Hispanics, the mean serum 25(OH)D was 48

nmol/L, and was inversely associated with %TBF (r = -0.28, P < 0.001) and TBF (r = -

0.33, P < 0.001) [188]. A study of Puerto Ricans in Puerto Rico showed that serum

25(OH)D was significantly inversely correlated with %TBF (r = -0.24, P=0.02) [60].

The aim of this cross-sectional study was to determine the association between

serum 25(OH)D concentration in relation to total body fat-mass (TBF) and percent-body-

fat (%TBF), and its various components, in the participants of the Boston Puerto Rican

144

Osteoporosis Study (BPROS). We included android fat-mass (AFM) and gynoid fat-mass

(GFM), using dual-energy x-ray absorptiometry (DXA).



Description of the recruitment and methods of the BPROS are described in detail

elsewhere [113]. Concisely, a bilingual interviewer conducted home visits and

administrated a detailed questionnaire on socioeconomic status, education, household

income, health and health behaviors that includes: medical diagnosis and current

medication use, depression, and psychological acculturation. Additionally, the

interviewer collected anthropometrics, physical performance measures, and measured

blood pressure. The day after this interview, the study phlebotomist collected 12-hr urine

and 12-hr fasting blood and saliva. Once the 2-yr follow up was complete, participants

were re-consented for the BPROS. By May 2013, 1257 participants completed 2year

follow-up visits and 970 had completed the BPROS study, which included a 24-hr urine

collection. All study protocols were approved by the Institutional Review Boards at Tufts

Medical Center and Northeastern University, and each participant provided written

informed consent.

There were no significant differences in socio-demographic characteristics

between those with complete or incomplete information. The present study includes

complete data for 700 Puerto Rican women and 270 Puerto Rican men.

145


For the purpose of this analysis, vitamin D deficiency was defined as serum

25(OH)D concentration below 50 nmol/L, insufficiency between 50-75 nmol/L, and

sufficiency above 75 nmol/L [26, 157]. Seasonality was defined as summer, July to

September; fall, October to December; winter, January to March; and spring, April to

June. Smoking status was categorized into yes or no for the current behavior. Vitamin D

from supplements was defined as: minimal 0-10 µg/day (200-400 IU/day); and high ≥ 10

µg/day (≥ 400 IU/day). Participant’s education was measured at baseline and defined

using the cutoff point of 8th grade while poverty was measured at each time point and

defined by the U.S. Department of Health and Human Services federal poverty levels.

Dietary vitamin D intake was estimated from a semi-quantitative food-frequency

questionnaire (FFQ) specifically designed for this population [91]. This FFQ has been

validated against serum carotenoids [116], and vitamin B12 [117] in a Hispanic

population aged ≥ 60 years. Those with energy intakes < 600 or > 4800 kilocalories

and/or > 10 questions blank on the FFQ were excluded from the analyses. Nutrient

intakes were calculated with the Nutrition Data System for Research (NDS-R) software

version 2007 (Nutrition Coordinating Center, University of Minnesota, Minneapolis,

MN, Food and Nutrient Database 2007, released 2007). Alcohol intake (g) was captured

by the FFQ.

146

Exposure Assessment:

All fat-mass measures were obtained using DXA whole-body scans (Lunar

Prodigy, GE Healthcare, Madison, WI) using software Encore version 6.1. Total body

scans were carried out on consenting participants, and all scans followed a standardized

protocol to ensure consistency. Daily quality assurance scans were performed, as well as

weekly scans of various thickness phantoms to examine the stability of the instrument.

The precision of the total body fat and lean tissue were, 0.94% and 0.77%, respectively

[191]. Android and gynoid measurements were based on the percentage of fat within each

body region. AFM (kg) was defined as the area between the ribs and the pelvis. The

upper demarcation is 20% of the distance between the iliac crest or the pelvis and the

femoral neck. The GFM (kg) was defined as the region including the hips and upper

thighs overlapping leg and trunk regions. Its upper demarcation, below the pelvis cut, is

defined as 1.5 times the android height, with total height as twice the height of the AFM

region (Figure 4.1). The lateral boundaries of both AFM and GFM were defined as arms

and legs [192]. Percentage total body fat, %TBF, is determined as a ratio of total-fat-mass

to total body mass, both in (kg).

Laboratory assays

Fasting blood samples (12-h) were drawn by a certified phlebotomist at the

participant’s home on the morning after the home interview, or as soon thereafter as

possible. Aliquots were saved and stored at -80°C until processed. Serum 25(OH)D

concentration was measured by extraction followed by 25I radioimmunoassay Packard

COBRA II Gamma Counter (DiaSorin Inc., Stillwater, MN 55082) catalog # 68100E

147

with intra-assay and inter-assay CV% of 10.8% and 9.4%, respectively. Serum intact

parathyroid hormone (iPTH) was detected by solid-phase, two-site chemiluminescent

enzyme-labeled immunometric assay using the Immulate 1000 Diagnostic Products

Corporation (Los Angeles, CA 90045-5597) with intra-assay and inter-assay CV% of

5.5%, and 7.9%, respectively. Serum creatinine was assessed with a colorimetric, kinetic

reaction using the Olympus AU400e with Olympus Creatinine Reagents (OSR6178)

(Olympus America Inc., Melville, NY) with intra-assay and inter-assay CV% of 2.0%,

and 4.0%, respectively.


All statistical analyses were conducted with SAS 9.1.3 (SAS Institute Inc, Cary,

NC); and all tests were 2-sided with P <0.05 considered statistically significant.

Dependent skewed variables were natural log-transformed to normalize their distributions

before the analysis. Demographic and clinical characteristics were compared using

Student’s t-test for continuous variables and chi square tests for categorical variables.

Characteristics were compared by sex using two-tailed t-test for two independent

samples, and with χ2 test for differences in proportion. Pearson partial linear correlation

analysis was used to examine the relationship between serum 25(OH)D concentration

with adiposity, adjusting for age and sex. We constructed general linear models (PROC

GLM) to model the cross-sectional associations between 25(OH)D (continuous) and fat-

mass (continuous). The Risk Factor model (RF-model) was adjusted for age (y), serum

creatinine (nmol/L), serum iPTH (pg/ml), and sex in the full model. The Dietary model

(D-model) included serum seasonality, dietary vitamin D and supplements, and total

148

energy (kcal/day) in addition to the prior model covariates. The Life-Style model (LS-

model) was adjusted for prior models in addition to physical activity, current smoking

status (y/n), alcohol intakes (g/d), poverty level, and education. For all linear models, we

checked the assumptions of linearity and homogeneity through examining the outcome

residuals against the exposure. For all final models, we performed influence diagnostics

tests to check for outliers and influential points.

RESULTS

The mean ages of women and men were 60.3 ± 7.5 and 59.5 ± 8.0, respectively

(Table 1). The mean serum 25(OH)D for women was 50.1 ± 19.0, significantly higher (P

= 0.01) compared to men, 47.0 ± 17.5 nmol/L. Women had significantly higher adiposity

than men for most measures. Women also had significantly higher dietary vitamin D and

supplement intakes. Men walked significantly more miles/day and had higher physical

activity scores compared to women. Men also reported higher levels of smoking and

alcohol consumption compared to the women.

Only AFM was significantly negatively correlated with serum 25(OH)D (r = -0.1,

P =0.02), while TBF approached significance (r = -0.06, P =0.07) (Table 4.3). The

multivariate regression analysis of the associations of DXA TBF-in kilogram-or %TBF-

were not significantly associated with serum 25(OH)D, adjusting for potential covariates

(Table 4.3).

In the RF-model, AFM was significantly negatively associated with serum

25(OH)D (β = -0.003, P =0.009). This inverse association remained significant after

further adjustment for the covariates (β = -0.003, P =0.01). Even though the interaction

149

term of AFM and sex was not significant, but regional adiposity might be sex-specific,

thus, we stratified by sex. Among Puerto Rican men, the additional adjustment for the

covariates of the LS-model did not attenuate the association between AFM and serum

25(OH)D (β = -0.002, P =0.008) explaining approximately 22% of the variability in

serum 25(OH)D. Among Puerto Rican women, only the additional adjustments of LS-

model covariates appreciably emphasized the significantly inverse association between

AFM and 25(OH)D (β = -0.003, P =0.03).

GFM was not a significant predictor of serum 25(OH)D when controlling for the

covariates of the RF-model. However, with additional adjustment for the covariates of D-

model and LS-model, GFM was positively significantly associated with 25(OH)D (β =

0.002, P =0.02). Interestingly, when the analyses were stratified by sex, the positive

associations remained significant among Puerto Rican women only (β = 0.002, P =0.04)

(Figure 4.2).

DISCUSSION

These results provide support to the current dialogue on adiposity, in general, and

fat-mass-in terms of amount and region-in particular, with vitamin D status. By initially

focusing on TBF-as measured by DXA-and serum 25(OH)D concentration, we did not

discern any significant association. Similarly, none of our models were significant for

%TBF. On the other hand, when we further investigated the role of regional adiposity,

precisely, android and gynoid fat-mass, our results showed a distinct picture. AFM was

negatively associated with serum 25(OH)D after adjusting for several potential covariates

150

including, age, serum creatinine, serum iPTH, seasonality, supplements and dietary

vitamin D, smoking, alcohol, and total energy, thus, demonstrating the independent

negative association of android adiposity on serum 25(OH)D concentration. In contrast,

GFM was significantly positively associated with serum 25(OH)D (P =0.02) independent

of age, serum creatinine, serum iPTH, seasonality, supplements and dietary vitamin D,

smoking, alcohol, and total energy.

We reported earlier that vitamin D deficiency remains prevalent in this Hispanic

subgroup, indicating that vitamin D deficiency is not a transient health concern, thus,

predisposing this population to higher levels of mortality and morbidity [182]. Recently,

we reported that the change in BMI over time was negatively associated with the change

in serum 25(OH)D concentration (P =0.01) independent of various known factors

including age, sex, serum creatinine, physical activity, smoking alcohol, seasonality,

poverty and education [193]. Taken together, our results are in agreement with a number

of epidemiological studies demonstrating that measures of adiposity including TBF, BMI,

and regional adiposity are inversely associated with serum 25(OH)D concentration [48,

49, 132, 171-173].

The association between vitamin D and adiposity, in general, and regional

adiposity, specifically, was explained by Lumb and colleagues in 1971. They were the

first to introduce the hypothesis of vitamin D sequestration in the adipose tissue and

muscle [76]. Later on, they substantiated that hypothesis and confirmed the link between

adiposity and vitamin D [77]. Subsequently, Wortsman and colleagues confirmed the

aforementioned premise, stating that deposition within the adipose tissue lowers the

bioavailability of serum 25(OH)D and 1,25(OH)2D leading to vitamin D

151

insufficiency/deficiency. They showed lower serum 25(OH)D and higher iPTH among

obese people [48]. AFM has been associated with higher mortality independent of BMI

[194] and is a known risk factor for many serious health disparities including CVD,

hypertension, dyslipidemia and insulin resistance [194-196], as well as inflammation

[179, 197].

The results of our study should be interpreted within the context of some

limitations. First, due to the cross-sectional nature, causality cannot be inferred, but the

results are in agreement with a number of cross-sectional studies [172, 188, 190].

Another potential limitation is that with the use of DXA, the differentiation between

visceral and subcutaneous fat is not feasible. However, it was shown by Savgan-Gurol

and colleagues that the percentage fat shown by DXA was significantly associated with

visceral fat (r = -0.83, P < 0.0001), while the association with subcutaneous was lower (r

= -0.77, P < 0.001) [198]. This only suggests that the importance of android fat-mass as

an indicator of central obesity might be greater compared to the relative distribution of

visceral and subcutaneous fat.

In conclusion, our findings of the negative, significant association of AFM on

serum 25(OH)D concentration; while GFM was positively associated with serum

25(OH)D concentration, highlights the importance of differences in body composition on

vitamin D status. Although it is clear that some of the association between regional

adiposity and serum 25(OH)D concentration may be attributed to the sequestration of

vitamin D in adipose tissues, it is still considered a complex relationship and more work

is needed to elucidate the relationship between adipose tissue and vitamin D and the

effect both to overall health.

152

BIBLIOGRAPHY

1. Mattei, J., et al., Allostatic load is associated with chronic conditions in the

Boston Puerto Rican Health Study. Social Science & Medicine, 2010. 70(12): p. 1988-96.



white neighborhood comparison group. J Gerontol A Biol Sci Med Sci, 2000. 55(2): p.

M90-7.

3. Tucker, K.L., O.I. Bermudez, and C. Castaneda, Type 2 diabetes is prevalent and

poorly controlled among Hispanic elders of Caribbean origin. Am J Public Health, 2000.

90(8): p. 1288-93.

4. Ogden CL, C.M., Kit BK, Flegal KM., Prevalence of obesity in the United States,

2009–2010. NCHS data brief, 2012, National Center for Health.

5. Flegal Km, C.M.D.K.B.K.O.C.L., Prevalence of obesity and trends in the

distribution of body mass index among us adults, 1999-2010. JAMA, 2012. 307(5): p.

491-497.

6. Flegal, K.M., et al., PRevalence and trends in obesity among us adults, 1999-

2008. JAMA, 2010. 303(3): p. 235-241.

7. Casas, Y.G., et al., Total and regional body composition across age in healthy

Hispanic and white women of similar socioeconomic status. The American Journal of

Clinical Nutrition, 2001. 73(1): p. 13-18.

8. Bermudez, O.I. and K.L. Tucker, Total and central obesity among elderly

Hispanics and the association with Type 2 diabetes. Obes Res, 2001. 9(8): p. 443-51.


mineral density in older Puerto Rican adults. The American Journal of Clinical Nutrition,

2011. 94(4): p. 1063-1070.

10. Wiklund, P., et al., Abdominal and Gynoid Fat Mass Are Associated with

Cardiovascular Risk Factors in Men and Women. Journal of Clinical Endocrinology &

Metabolism, 2008. 93(11): p. 4360-4366.

153

11. McCarty, M.F., A paradox resolved: the postprandial model of insulin resistance

explains why gynoid adiposity appears to be protective. Medical Hypotheses, 2003.

61(2): p. 173-176.

12. Mithal, A., et al., Global vitamin D status and determinants of hypovitaminosis D.


13. Holick, M.F. and T.C. Chen, Vitamin D deficiency: a worldwide problem with

health consequences. Am J Clin Nutr, 2008. 87(4): p. 1080S-6S.

14. Holick, M.F., Vitamin D status: measurement, interpretation, and clinical

application. Ann Epidemiol, 2009. 19(2): p. 73-8.

15. Liel, Y., et al., Low circulating vitamin D in obesity. Calcified tissue international,

1988. 43(4): p. 199-201.

16. Weishaar, T. and J.M. Vergili, Vitamin D Status Is a Biological Determinant of

Health Disparities. Journal of the Academy of Nutrition and Dietetics, 2013. 113(5): p.

643-651.

17. Young, K.A., et al., Association of Plasma Vitamin D Levels with Adiposity in

Hispanic and African Americans. Journal of Clinical Endocrinology & Metabolism,

2009. 94(9): p. 3306-3313.

18. Jamal-Allial, A.A. and K.L. Tucker, Serum Vitamin D Concentrations, Dietary

Sources, and its relation to BMI among Puerto Ricans Living in the Greater Boston Area.

To be submitted, 2013.

19. Renzaho, A.M., J.A. Halliday, and C. Nowson, Vitamin D, obesity, and obesity-

related chronic disease among ethnic minorities: a systematic review. Nutrition, 2011.

27(9): p. 868-79.

20. Norman, A.W., Minireview: vitamin D receptor: new assignments for an already

busy receptor. Endocrinology, 2006. 147(12): p. 5542-8.

21. Tuohimaa, P., Vitamin D and aging. Journal of Steroid Biochemistry & Molecular

Biology, 2009. 114(1-2): p. 78-84.

22. Brock, K., et al., Low vitamin D status is associated with physical inactivity,

obesity and low vitamin D intake in a large US sample of healthy middle-aged men and

women. J Steroid Biochem Mol Biol, 2010. 121(1-2): p. 462-6.

154

23. McKinney, K., C.R. Breitkopf, and A.B. Berenson, Association of race, body fat

and season with vitamin D status among young women: a cross-sectional study. Clinical

Endocrinology, 2008. 69(4): p. 535-541.

24. Wortsman, J., et al., Decreased bioavailability of vitamin D in obesity. The


25. Snijder, M.B., et al., Adiposity in relation to vitamin D status and parathyroid

hormone levels: a population-based study in older men and women. Journal of Clinical

Endocrinology & Metabolism, 2005. 90(7): p. 4119-23.

26. Parikh, S.J., et al., The Relationship between Obesity and Serum 1,25-Dihydroxy

Vitamin D Concentrations in Healthy Adults. Journal of Clinical Endocrinology &

Metabolism, 2004. 89(3): p. 1196-1199.

27. Palacios, C., et al., Determinants of vitamin D status among overweight and obese

Puerto Rican adults. Ann Nutr Metab, 2012. 60(1): p. 35-43.


hydroxyvitamin D and body mass index: the Tromsø study. European Journal of

Nutrition, 2010. 49(7): p. 401-407.

29. Botella-Carretero, J.I., et al., Vitamin D deficiency is associated with the

metabolic syndrome in morbid obesity. Clinical Nutrition, 2007. 26(5): p. 573-80.

30. Arunabh, S., et al., Body Fat Content and 25-Hydroxyvitamin D Levels in Healthy

Women. Journal of Clinical Endocrinology & Metabolism, 2003. 88(1): p. 157-161.

31. Dawson-Hughes, B., et al., Estimates of optimal vitamin D status. Osteoporosis

International, 2005. 16(7): p. 713-716.



Puerto Rican and non-Hispanic adults. American Journal of Epidemiology, 1998.

148(5): p. 507 - 518.

34. Bermudez, O., et al., Hispanic and non-Hispanic white elders from Massachusetts

have different patterns of carotenoid intake and plasma concentrations. Journal of

Nutrition, 2005. 135(6): p. 1496 - 1502.

155

35. Kwan, L.L., O.I. Bermudez, and K.L. Tucker, Low vitamin B-12 intake and status

are more prevalent in Hispanic older adults of Caribbean origin than in neighborhood-

matched non-Hispanic whites. Journal of Nutrition, 2002. 132(7): p. 2059-64.

36. White, J., et al., Precision of single vs bilateral hip bone mineral density scans.

Journal of clinical densitometry : the official journal of the International Society for

Clinical Densitometry, 2003. 6(2): p. 159-62.

37. Mazess, R.B., et al., Dual-energy x-ray absorptiometry for total-body and

regional bone-mineral and soft-tissue composition. The American Journal of Clinical

Nutrition, 1990. 51(6): p. 1106-12.

38. Forrest, K.Y.Z. and W.L. Stuhldreher, Prevalence and correlates of vitamin D

deficiency in US adults. Nutrition Research, 2011. 31(1): p. 48-54.

39. Jamal-Allial, A.A., J.L. Griffith, and K.L. Tucker, The Longitudinal Association

of Vitamin D Serum Concentrations & Adiposity Phenotypes. To be submitted, 2013.

40. Renzaho, A.M.N., J.A. Halliday, and C. Nowson, Vitamin D, obesity, and obesity-

related chronic disease among ethnic minorities: A systematic review. Nutrition, 2011.

27(9): p. 868-879.

41. Lumb, G.A., E.B. Mawer, and S.W. Stanbury, The apparent vitamin D resistance

of chronic renal failure: A study of the physiology of vitamin D in man. The American

Journal of Medicine, 1971. 50(4): p. 421-441.

42. Mawer, E.B., et al., The distribution and storage of vitamin D and its metabolites

in human tissues. Clin Sci, 1972. 43(3): p. 413-31.

43. Jacobs, E.J., et al., WAist circumference and all-cause mortality in a large us

cohort. Archives of Internal Medicine, 2010. 170(15): p. 1293-1301.

44. Panagiotakos, D.B., et al., The implication of obesity and central fat on markers

of chronic inflammation: The ATTICA study. Atherosclerosis, 2005. 183(2): p. 308-315.

45. Dos Santos, R.E., et al., Relationship of body fat distribution by waist

circumference, dual-energy X-ray absorptiometry and ultrasonography to insulin

resistance by homeostasis model assessment and lipid profile in obese and non-obese

postmenopausal women. Gynecological Endocrinology, 2005. 21(5): p. 295-301.

156

46. Compher, C. and K.O. Badellino, Obesity and Inflammation: Lessons From

Bariatric Surgery. Journal of Parenteral and Enteral Nutrition, 2008. 32(6): p. 645-647.

47. Fontana, L., et al., Visceral Fat Adipokine Secretion Is Associated With Systemic

Inflammation in Obese Humans. Diabetes, 2007. 56(4): p. 1010-1013.

48. Savgan-Gurol, E., et al., Waist to hip ratio and trunk to extremity fat (DXA) are

better surrogates for IMCL and for visceral fat respectively than for subcutaneous fat in

adolescent girls. Nutrition & Metabolism, 2010. 7(1): p. 86.

49. Dietary Reference Intakes for Calcium and Vitamin D. Committee to Review

Dietary Reference Intakes for Vitamin D and Calcium. , in 2010, Institute of Medicine:

Washington, DC: National Academy of Sciences.

157

FIGURE 4.1 The illustrations of android region (AFM) and gynoid region (GFM)

158

FIGURE 4.2 The β coefficients of the association between serum 25(OH)D concentration with gynoid and android fat among Puerto Rican men and women adjusted for age and serum creatinine (mg/dL), serum iPTH (pg/ml), total fat-mass (g), seasonality, dietary vitamin D and supplements, total energy (kcal/day), physical activity, alcohol, smoking, poverty level, education and sex in the All data.

159

TABLE 4.1 Descriptive characteristics of men and women in Boston Puerto Rican

Osteoporosis Study (BPROS)

Variables Women (n=699 )

Men (n=273 )

P-value

Age (y) 60.3 ± 7.5 59.5 ± 8.0 0.2

Education (above 8th grade %) 49.0 44.3 0.1

Poverty (%) 59.0 49.2 0.009

Serum 25(OH)D (nmol/L) 50.1 ± 19.0 47.0 ± 17.5 0.01

Adiposity Variables

BMI (kg/m2) 33.0 ± 7.0 30.2 ± 5.5 <0.0001

Waist Circumference (cm) 104.0 ± 15.0 104.3 ± 14.3 0.4

Total Fat-mass (kg) 36.0 ± 1.5 26.5 ± 1.1 <0.0001

Android Fat-mass (kg) 3.4 ± 1.3 3.0 ± 1.4 0.0007

Gynoid Fat-mass (kg) 6.0 ± 2.1 4.1± 2.0 <0.0001

Andriod:Gynoid 1.05 ± 0.13 1.3 ± 0.22 <0.0001

Average fat percentage 44.4 ± 6.5 30.0± 7.5 <0.0001

Other Covariates

Current- smoking (%) 18.0 30.0 <0.0001

Current- alcohol drinking (%) 36.0 51.0 <0.0001

Physical Activity Score$‡ 31.3 ± 4.0 33.0 ± 6.0 <0.0001

Miles walking/day$‡ 1.4 ± 2.0 2.0 ± 2.4 0.0002

160

Dietary Measures

Vitamin D supplement (>200IU/day) # (%) 30.2 18.5 0.0002

Total vitamin D intake (IU/day) || 287.5 ± 152.0 257.0 ± 123.0 0.002

Total energy intake (kcal/day) 1770.0 ± 825 2250± 1200 <0.0001

BMI, body mass index; 25(OH)D, 25-hydroxyvitamin D. 1Data presented as mean ± SD for continuous variables, or as % for categorical variables $Geometric mean ± SD For between sex’s comparisons, t-test was used for means and Chi-square test χ2- test was used for percentages. ‡ Range for Physical Activity Score = 35.5 for men and 26.4 for women. Range for Miles/day = 22 for men and 20 miles/day for women. #The recommended dietary allowance (RDA) for vitamin D for men and women aged 19-70 years =15 ug /day [82] ||Adjusted for energy intake using ANOVA

161

TABLE 4.2 Pearson partial linear correlation of serum 25(OH)D concentration and

adiposity phenotype at 2-year

2-year

25(OH)D p-value

TBF - 0.06 0.07

%TBF - 0.03 0.3

AFM -0.1 0.02

GFM -0.04 0.2

Pearson partial linear correlation coefficient (r) is age and sex adjusted Abbreviations: BMI, body mass index; 25(OH)D, 25-hydroxyvitamin D; WC, waist circumference; AFM, abdominal fat-mass; TBF, total body fat; %TBF, total body fat percentage

162

TABLE 4.3 Regression coefficients (± SE) from cross-sectional regression analysis of

TBF and %TBF on plasma 25(OH)D concentration.


Adiposity Phenotype

TBF

%TBF

β SE P-value β SE P-value

Risk Factor Model

- 0.0001 0.00005 0.2 - 0.03 0.1 0.7

Dietary Model

- 0.00005 0.00005 0.2 - 0.03 0.1 0.6

Life-Style Model

- 0.0001 0.00005 0.1 - 0.1 0.1 0.3

TBF, total body fat; TBF%, total body fat percentage Risk Factor Model: adjusted for age and creatinine (mg/dL), serum iPTH (pg/ml) and Dietary Model: adjusted for model 1 variables plus seasonality, dietary vitamin D and supplements, total energy (kcal/day).

Life-Style Model: adjusted for model 1+ 2 variables plus physical activity, alcohol, smoking, poverty level and education.

163

TABLE 4.4 Regression coefficients (± SE) from cross-sectional regression analysis of AFM and GFM on plasma 25(OH)D

concentration.

Adiposity Phenotype


AFM β ± (SE)

GFM β ± (SE)

All1 Men Women All1 Men Women

β ± SE P β ± SE P β ± SE P β ± SE P β ± SE P β ± SE P

Risk Factor Model

-0.003 ± 0.001

0.009 -0.004 ± 0.003

0.1 -0.002 ± 0.001

0.1 0.001 ± 0.001

0.06 0.001 ± 0.002

0.1 0.002 ± 0.001

0.09

Dietary Model:

-0.003 ± 0.001

0.01 -0.002 ± 0.001

0.008 -0.001 ± 0.0005

0.2 0.002 ± 0.001

0.03 0.001 ± 0.002

0.4 0.002 ± 0.001

0.04

Life-Style Model

-0.003 ± 0.001

0.01 -0.002 ± 0.001

0.01 -0.003 ± 0.001

0.03 0.002 ± 0.001

0.02 0.001 ± 0.002

0.5 0.002 ± 0.001

0.04

AFM, android- fat-mass; GFM, gynoid-fat-mass. Risk Factor Model: adjusted for age and serum creatinine (mg/dL), serum iPTH (pg/ml), total fat-mass (g) and sex in the case of the combined data1 Dietary Model: adjusted for model 1 variables plus seasonality, dietary vitamin D and supplements, total energy (kcal/day).

Life-Style Model: adjusted for model 1+ 2 variables plus physical activity, alcohol, smoking, poverty level, education.

164

CHAPTER SIX

DISCUSSION AND SUMMARY

Understanding the association between plasma 25-hydroxyvitamin D concentration

(25(OH)D) concentration and cardiovascular disease (CVD) is important. . This dissertation was

designed to examine the multiple pathways through which serum 25(OH)D as the biomarker of

vitamin D status could affect longitudinally as well as cross-sectionally a number of CVD

associated risk factors, namely, lipid profile, adiposity, and inflammation among the participants

of the Boston Puerto Rican Health Study (BPRHS) and the Boston Puerto Rican Osteoporosis

Study (BPROS).

Foremost, it is well documented in the literature that dyslipidemia, i.e. high plasma LDL-

C, TG and low plasma HDL-C, is highly associated with increased risk of CVD along with other

factors such as high blood pressure [2, 128, 199]. In the first report of this doctoral work, our

results are in apparent contrast to scores of reports that indicate serum 25(OH)D concentration is

positively associated with plasma high-density lipoprotein (HDL-C), and negatively associated

with plasma low-density lipoprotein (LDL-C), total cholesterol (TC), and triglycerides (TG) [4-

10]. We showed that even though serum 25(OH)D concentration was positively associated with

plasma HDL-C cross-sectionally, the longitudinal analyses illustrated different yet interesting

associations; it showed that baseline-25(OH)D demonstrated a negative, yet significant,

association with 2.5-year change in plasma HDL-C concentration (β = -0. 04, P =0.01); an

association independent of important putative confounders including age, sex, body mass index

(BMI), total energy, physical activity, smoking, alcohol, seasonality, poverty and education. We

165

demonstrated this association not only within the life-style model, but throughout the

confounding by indication adjustment in our CVD-model. Similar relationships were found with

plasma TG, but not with LDL-C and TC. In the prospective Health Professionals Follow-up

Study (HPFS), Giovannucci and colleagues demonstrated a positive linear trend of LDL-C,

HDL-C and TC, as well as a negative trend with plasma TG across serum 25(OH)D quartiles

among 900 US men, ages 64 and older [125]. Another study showed that serum 25(OH)D was

independently associated with higher HDL-C and lower TG; but not TC [107]. Interestingly, two

studies reported that both HDL-C and LDL-C were positively associated with serum 25(OH)D

concentration, thus, both are being affected in a the same direction, i.e. increasing their plasma

concentration. It is well known that one favorable approach to lower CVD risk is to increase

HDL-C and lower LDL-C [2, 128]. A report by Melamed and colleagues, found that when

analyzing Third National Health and Nutrition Examination Survey (NHANES III), serum

25(OH)D concentration were significantly inversely associated with TC, among 13,331 US men

and women ages 20 years and older. However, they found no association between plasma HDL-

C and serum 25(OH)D among their participants [96]. Another longitudinal study with a 14 year

follow up found that among 1762 participants, serum 25(OH)D and plasma TG were inversely

associated [104]. Few key differences can be outlined and prior studies are mostly cross-

sectional; no proof of a causal relationship can be inferred. One difference to illustrate is that our

results are longitudinal, and we have controlled for numerous confounders, which were not

necessarily available among previous cross-sectional studies.

On the contrary our results are in accordance with a number of reports that found serum

25(OH)D concentration might not be associated with atheroprotective lipid profile [96, 104, 105,

123, 127]. A study by Jorde and colleagues found that TC was significantly associated with

166

increased serum 25(OH)D concentration among 8,018 Norwegian men and women, ages 43-70

years [104]. Two other studies found that plasma TG were positively associated with serum

25(OH)D; however, both studies investigated this effect among children [123, 127]. Intervention

studies in the form of placebo-controlled randomized trials (RCT’s) are designed to confirm or

dispute any causal relationship. Exceptionally few RCT have been conducted to investigate the

question in discussion, and their results have largely been inconsistent on the effect of vitamin D

supplementations on improving lipid profile [108, 129-131]. A study by Zittermann and

colleagues showed that daily vitamin D supplementation (3320 IU/day) for one year had a

significant effect on plasma lipids. They found that the supplemented group had lower plasma

TG compared to the placebo, -13.5% and +3% (P =0.001) correspondingly; nonetheless, the

supplemented group exhibited a significant increase in plasma LDL-C [129]. Conversely, this

was not the case for the study by Jorde and colleagues; they were not able to demonstrate that

one year of vitamin D supplementation, 40, 000 IU/week, improved the plasma lipids of the 438

overweight/obese participants [108]. Another study showed that vitamin D supplementation

might not be atheroprotective among postmenopausal women [130]. Most of these RCT’s can be

criticized for a rather low sample size, which might have contributed to the null results.

The relationship between plasma HDL-C and serum 25(OH)D concentration is

particularly noteworthy. The significance of the interaction term of smoking status with serum

25(OH)D was interesting. The significant impact of serum 25(OH)D on HDL-C among non-

smokers was unexpected, however similar results were reported by Jorde and colleagues [132].

They found that among the participants of the Tromsø study, smokers had higher levels of serum

25(OH)D concentration; about 15-20%, compared to non-smokers. They thought it might be due

to a difference in methodology, i.e. immunological vs. LC-MSMS, and excluded them from the

167

analysis. The pathophysiological mechanism by which smoking may impact serum 25(OH)D

concentration or the metabolism of 1,25(OH)2D has not been investigated. However, it has been

discussed that smoking might influence 25-hydroxylase (CYP2R1) thus, lowering serum

25(OH)D concentration [133]. Another point that might explain this paradox is that, smoking

contains nicotine, which exerts anti-inflammatory properties [134, 135].

Despite the disagreement of our results with many epidemiologic and ecological studies

for the protective effective role of serum 25(OH)D on lipid profile as a major risk factor for

CVD, the contradictory yet significant evidence we show cannot be disregarded. One

explanation for these phenomena is that vitamin D induces the intestinal absorption of dietary

calcium, thus lowering calcium fecal excretion [37, 95]. This might affect the amount of fatty

acids in the fecal due to the reduction in the formation of insoluble fatty soaps [136]. Thus,

leading to an increased rate of the absorption of fat, thus higher plasma LDL-C and TG

concentration, in particular, and TC in general [137]. In addition, higher vitamin D might

decrease cholesterol excretion because of the lower availability of calcium within the intestine,

and hence less binding to bile acids [138]; this would translate into less cholesterol converting

into bile acids. One study suggested that higher calcium intake might augment the prior

phenomena and lower LDL-C by increasing not only vitamin D but it has been accompanied by

calcium supplements. A 15-week weight loss trial study by Major and colleagues showed that

among 63 healthy, overweight/obese women, the daily supplementation of 600 mg of calcium

and 200 IU of vitamin D significantly improved LDL-C as well as LDL: HDL-C ratio, while

HDL-C and TG approached significance level [139].

Secondly, there is considerable evidence that inflammation is one element of the

development of CVD. Higher serum concentration of inflammatory markers such TNF-α, IL-6

168

and hsCRP have emerged as independent risk factors for CVD [165, 166]. Therefore, attenuating

these inflammatory markers concentration may influence risk of such illnesses. Vitamin D may

have the ability to attenuate the high levels of inflammatory markers by down-regulating a

number of genes [37-39, 144].

A second aim of this doctoral work was to study the cross-sectional association between

serum 25(OH)D and inflammatory markers, namely, interleukin 6 (IL-6) and tumor necrosis

factor-alpha (TNF-α) as well as high-sensitivity C-reactive protein (hsCRP) among the

participants of the BPROS. Our data suggest that serum 25(OH)D concentration explained about

6-20% of the variation in the IL-6 and TNF-α concentration, respectively. Our data provides

evidence of an inverse association between serum 25(OH)D and inflammatory markers, namely,

IL-6 and TNF-α, among the participants of the BPROS; irrespective of age, sex, seasonality,

adiposity phenotype and lifestyle factors that include smoking, physical activity and alcohol

consumption. This population, as reported before, exhibits a high prevalence of vitamin D

deficiency [159]. Our findings from the multivariate regression analysis, especially for IL-6 and

TNF-α, showed that irrespective of the various constructed models, serum 25(OH)D

concentration remained a significant predictor of IL-6 and TNF-α, but not of hsCRP, after

controlling for potential covariates. These results are in accordance with other reports on the

association between serum 25(OH)D concentration and inflammatory markers [98, 154, 155]. In

further support, our results are consistent with in vitro and animal studies that showed TNF-α

suppression by vitamin D [97, 152, 160, 161] and IL-6 [97, 147, 162]. A RCT by Schleithoff and

Colleagues, demonstrated that 9 months of daily supplementation of 2000 IU of vitamin D and

500 mg of calcium significantly lowered serum TNF-α concentration compared to baseline

169

among congestive heart failure participants [153]. In addition, another RCT demonstrated an

inhibition of TNF-α concentration with 3332 IU of vitamin D daily [129].

The lack of association between serum 25(OH)D and hsCRP is not in concurrence with a

number of cross-sectional reports [98, 163]; there are a number of plausible explanations. First,

hsCRP concentration can be affected by acute phase reactant due to viral infection. However,

that might not be the case with the presence of constant yet potent stimuli [149]. Another point to

emphasize is that this specific population has a high prevalence of health disparities that includes

metabolic syndrome, depression, and diabetes, which might mask the association [109, 164].

Additionally, in this population the mean of 25(OH)D concentration reaches only to 50 nmol/L;

indicating widespread vitamin D deficiency and thus, very limited distribution and a very low

number of sufficient levels to show significant differences. In our analysis, the adjustment of

white blood cell counts, which is an indicator of current infection, did not improve the null

association between hsCRP and serum 25(OH)D concentration.

Thirdly, adiposity increases the risk of numerous chronic diseases such as CVD and

type-2-diabet among others [200, 201]. We intended for our third report of this doctoral work to

investigate the magnitude of serum 25(OH)D concentration being affect by adiposity phenotype

over time. Our results confirm an inverse association between low serum 25(OH)D concentration

and adiposity phenotype i.e. body mass index (BMI) and waist circumference (WC). Our

observation of lower adjusted mean serum 25(OH)D concentration in obese, severely obese and

morbidly obese participants is in agreement with many studies [132, 172, 176, 177]. Baseline-

BMI was negatively associated with the change in serum 25(OH)D concentration (P =0.01),

independent of known factors including age, sex, serum creatinine, physical activity, smoking,

alcohol, seasonality, poverty and education. We adjusted for covariates in three different models,

170

all of which showed significant longitudinal negative association. Conversely, we did not

observe similar associations with WC, even though it was significantly correlated with serum

25(OH)D at both time points.

A mechanism that can explain the phenomena of low serum 25(OH)D and increased

adiposity risk is not clear. The review by Earthman and colleague discusses number of proposed

theories that link Vitamin D deficiency to adiposity [178]. One of the theories suggests that since

Vitamin D is fat-soluble it might be sequestered within adipose tissue. They further suggest that

the adipose tissue of obese individuals may have higher 25(OH)D uptake and storage relative to

lean individuals [48, 178]. An additional theory suggests that due to the fact that adiposity,

especially visceral adiposity, is considered a high inflammatory state, this contributes to lower

serum 25(OH)D concentration [178, 179]. Finally, with increased adiposity, there may be an

increase of the action of 24-hydroxylase, thus, leading to amplified catabolism of Vitamin D

[178, 180]. Consequently, further research is needed to further elucidate our understanding of

how vitamin D might be affected by adiposity or even how vitamin D affects obesity itself.

The majority of the participants of BPROS- more than 60%- had serum 25(OH)D

concentration below 50 nmol/L, which is considered a deficiency according to the Endocrine

Society [115]. The most recent report by the National Health and Nutrition Examination Survey

(NHASES) III, showed that more than 42% of their 2001-2006 participants had vitamin D below

50 nmol/L; when stratified by ethnicity, Hispanics were second highest in vitamin D deficiency

with 70% [181]. Our prior report indicated that among 1502 participants of the BPRHS, more

than 70% had serum 25(OH)D concentration below 50 nmol/L at baseline, concurring with the

NHANES III data [159]. This greater risk of vitamin D deficiency may be related to greater risk

of health disparities [182]. A report by Palacios and colleagues showed that Puerto Ricans who

171

are residing in Puerto Rico, had higher serum 25(OH)D concentration, approximately 75 nmol/L,

compared to our population, at baseline 43.4 ± 17.2 nmol/L [159] and 2yr 50.0 ± 19.0 nmol/L.

One theory to illustrate here is that this difference in serum 25(OH)D concentration might be due

to sun exposure, generally, and latitude, specifically, i.e. PR’s latitude is 18.26° while Boston’

latitude is 42.35° [26].

Lastly, we aspired in our fourth report of this doctoral work to not only explore the

association between serum 25(OH)D concentration and adiposity phenotype, in general, but

rather to investigate in more depth that association total as well as regional fat mass. Our report

initially focused on total body fat (TBF)-as measured by dual-energy x-ray absorptiometry

(DXA) and serum 25(OH)D concentration. Similarly, percent body fat (%TBF) was not

significantly associated with 25(OH)D. On the other hand, when we further investigated the role

of regional adiposity, android (AFM) and gynoid fat-mass (GFM), results differed. AFM, was

negatively associated with serum 25(OH)D after adjusting for several potential covariates

including, age, serum creatinine, serum iPTH, seasonality, supplements and dietary vitamin D,

smoking, alcohol, and total energy. On the other hand, GFM was significantly positively

associated with serum 25(OH)D (P =0.02) independent of age, serum creatinine, serum PTH,

seasonality, supplements and dietary vitamin D, smoking, alcohol, and total energy.

We reported earlier that vitamin D deficiency remains prevalent in this Hispanic

subgroup, indicating that vitamin D deficiency is not a transient health concern, thus,

predisposing this population to higher levels of mortality and morbidity [182]. Recently, we

reported that the change in BMI over time was negatively associated with the change in serum

25(OH)D concentration (P =0.01) independent of various known factors including age, sex,

serum creatinine, physical activity, smoking alcohol, seasonality, poverty and education [193].

172

Taken together, our results are in agreement with a number of epidemiological studies

demonstrating that measures of adiposity including TBF, BMI, and regional adiposity are

inversely associated with serum 25(OH)D concentration [48, 49, 132, 171-173].

Our findings should be interpreted within the context of a few limitations. First, our

population is primarily Puerto Ricans, therefore our findings could be very specific to this

particular admixture population, with genetic variation interacting not only with the

environment, but social factors as well [140-143]; thus, generalizability may not be an option.

Additionally, even though we have controlled for all the known potential covariates, residual

confounding is still be a possibility.

Some of our findings came from our cross-sectional study; hence, it provides neither

further insights about causation nor long-term status. Longitudinal studies with two or more time

points are ideal alternative to investigate the role of serum 25(OH)D concentration on CVD risk

factors; because serum 25(OH)D concentration can be assessed before the outcome, thus,

allowing us to avoid any potential bias. Nevertheless, our findings show very unique results to

the Puerto Rican population. Another limitation is that we have no direct measure of sun

exposure at baseline. Another potential limitation is that with the use of DXA, the differentiation

between visceral and subcutaneous fat is not feasible. However, it was shown by Savgan-Gurol

and colleagues that the percentage fat shown by DXA was significantly associated with visceral

fat (r = -0.83, P < 0.0001), while the association with subcutaneous was lower (r = -0.77, P <

0.001) [198]. This only suggests that the importance of android fat-mass as an indicator of

central obesity might be greater compared to the relative distribution of visceral and

subcutaneous fat. Another potential bias that might affect our results is selection bias that might

have occurred in our study. Even though selection bias may not be possible to appropriately

173

evaluate how it may affected our results, it is possible that those who declined to participate in

our study may have a higher socioeconomic status, as well as highly educated, thus, may higher

probability of having sufficient serum 25(OH)D concentration [187]

In conclusion, our longitudinal analysis showed that higher 25(OH)D serum

concentration did not translate into an increase of the atheroprotective plasma HDL-C, nor a

decrease of the un-atheroprotective plasma TG in this population of older Puerto Rican adults

living in the Boston area. Moreover, our results showed that smoking might be an effect modifier

of the longitudinal association of serum 25(OH)D and HDL-C, as well as LDL-C. Granted our

work should be replicated in other populations, our findings provide opposing yet significant

evidence of the effect of vitamin D on plasma lipid profile.

Secondly, our study supports the hypothesis that inflammatory markers such as IL-6 and

TNF-α are inversely associated with serum 25(OH)D concentration among the participants of the

BPROS. These findings support the need to re-examine the determination of optimal vitamin D

status. However, more studies, especially longitudinal prospective studies, are needed to fully

characterize the relationship between serum 25(OH)D concentration and inflammatory markers.

Thirdly, serum 25(OH)D concentration was inversely associated with 2.5-yr change in

BMI. The adjusted-means of serum 25(OH)D concentration showed declining trend, at both time

points, with the increase in BMI especially when reaching class III obesity. Our findings of the

negative significant association of AFM on serum 25(OH)D concentration provides compelling

evidence that AFM may affect serum 25(OH)D concentration; while GFM was positively

associated with serum 25(OH)D concentration, Even though it is clear that some segments of the

regional adiposity and serum 25(OH)D concentration relationships may be attributed to the

sequestration of vitamin D in adipose tissues, it is still considered a complex relationship and

174

more work is needed to elucidate the relationship between adipose tissue and vitamin D and the

effect both to overall health.

Taken together, it appears that serum 25(OH)D concentration is strongly influenced by

adiposity, which is considered a chronic low-grade inflammatory condition, thus, by sequestering

25(OH)D within the adipose tissue, neither the high inflammatory condition might be

ameliorated nor dyslipidemia could be improved by vitamin D. Figure 5.1 summarize the

findings of this dissertation work.

Vitamin D deficiency/insufficiency has emerged as an autonomous risk factor of

cardiovascular mortality. Thus, the need for an enhanced understanding of the epidemiological

data as well as the pathophysiological pathways by which this pleiotropic hormone may or may

not affect cardiovascular health is imperative. This understanding will result in better informing

of future clinical trials and ultimately results in improved evidence-based therapeutic

recommendations. Research findings by Bidchoff-Ferrari and Colleagues estimated that 1000

IU/day of vitamin D supplements 50% of the younger and older adults to sufficient levels of

serum 25(OH)D concentration, i.e. 75 nmol/L. However, it will take twice as much of that,

meaning 2000 IU/day to bring 85-90% of the population to those sufficient concentration [202].

Thus, for ethnic minorities who are facing higher health disparities may require 2000 IU/day of

vitamin D supplement. Despite the fact that it is premature to make such recommendations for

the rationale of precluding CVD risk, achieving sufficient concentration of 25(OH)D is essential.

Even though there is plethora of scientific information of vitamin D and its receptor, there is an

urgent necessity to further understand and elucidate the biochemical and transcriptional effect of

vitamin D and its various analogs on CVD, as well as to search for new targets to treat CVD and

ultimately reduce health disparities.

175

Notwithstanding the above limitations, our study has substantial strengths: a large sample

size with low attrition ; we have complete data on 1504 Puerto Rican men and women at

baseline, 1257 at 2-year and 970 within the ancillary bone study. This large sample size is ample

to detect differences with minimally 90% power. In addition, we have used detailed

questionnaires along with an FFQ that allowed detailed information.

In summary, our findings suggest that risk factors of CVD are highly associated with

lower 25(OH)D concentration among Puerto Rican men and women. Results from this

dissertation work may provide guidance for future interventional studies for an understudied yet

fast growing population that appears to have high health disparities. Improving vitamin D status

may not solve all of their health disparities; however, it might lower their risk of developing

chronic disease.

176

FIGURE 5.1 Diagram illustrates the relationship between serum 25(OH)D and the various CVD risk factors among the participants of the BPRHS

177

BIBLIOGRAPHY

1. Roger, V.L., et al., Heart disease and stroke statistics--2011 update: a report

from the American Heart Association. Circulation, 2011. 123(4): p. e18-e209.

2. Bonow, R.O., Primary Prevention of Cardiovascular Disease. Circulation, 2002.

106(25): p. 3140-3141.

3. Balagopal, P., et al., Nontraditional Risk Factors and Biomarkers for Cardiovascular

Disease: Mechanistic, Research, and Clinical Considerations for Youth. Circulation,

2011. 123(23): p. 2749-2769.

4. Giovannucci, E., et al., 25-hydroxyvitamin d and risk of myocardial infarction in men: A

prospective study. Archives of Internal Medicine, 2008. 168(11): p. 1174-1180.

5. Maki, K.C., et al., Serum 25-hydroxyvitamin D is independently associated with high-

density lipoprotein cholesterol and the metabolic syndrome in men and women. Journal

of Clinical Lipidology, 2009. 3(4): p. 289-296.

6. Melamed, M.L., et al., 25-hydroxyvitamin d levels and the risk of mortality in the general

population. Archives of Internal Medicine, 2008. 168(15): p. 1629-1637.

7. Jorde, R., et al., High serum 25-hydroxyvitamin D concentrations are associated with a

favorable serum lipid profile. European Journal of Clinical Nutrition, 2010. 64(12): p.

1457-64.

8. Delvin, E.E., et al., Vitamin D Status Is Modestly Associated with Glycemia and

Indicators of Lipid Metabolism in French-Canadian Children and Adolescents. The

Journal of Nutrition, 2010. 140(5): p. 987-991.

9. Williams, D.M., A. Fraser, and D.A. Lawlor, Associations of vitamin D, parathyroid

hormone and calcium with cardiovascular risk factors in US adolescents. Heart, 2011.

97(4): p. 315-320.

10. Karhapää, P., et al., Diverse associations of 25-hydroxyvitamin D and 1,25-dihydroxy-

vitamin D with dyslipidaemias. Journal of Internal Medicine, 2010. 268(6): p. 604-610.


weight loss on cardiovascular disease risk markers. The American Journal of Clinical

Nutrition, 2009. 89(5): p. 1321-1327.

178

12. Heikkinen, A., et al., Long-term vitamin D3 supplementation may have adverse effects on

serum lipids during postmenopausal hormone replacement therapy. European Journal of

Endocrinology, 1997. 137(5): p. 495-502.

13. Jorde, R., et al., No improvement in cardiovascular risk factors in overweight and obese

subjects after supplementation with vitamin D3 for 1 year. Journal of Internal Medicine,

2010. 267(5): p. 462-472.

14. Nagpal, J., J.N. Pande, and A. Bhartia, A double-blind, randomized, placebo-controlled

trial of the short-term effect of vitamin D3 supplementation on insulin sensitivity in

apparently healthy, middle-aged, centrally obese men. Diabetic Medicine, 2009. 26(1): p.

19-27.


hydroxyvitamin D and body mass index: the Tromsø study. European Journal of

Nutrition, 2010. 49(7): p. 401-407.

16. Yoon, V., N.M. Maalouf, and K. Sakhaee, The effects of smoking on bone metabolism.


17. Lacour, B., et al., Parathyroid function and lipid metabolism in the rat. Miner Electrolyte

Metab, 1982. 7(3): p. 157-65.

18. Deeb, K.K., D.L. Trump, and C.S. Johnson, Vitamin D signalling pathways in cancer:

potential for anticancer therapeutics. Nature reviews. Cancer, 2007. 7(9): p. 684-700.

19. Reid, I.R., Effects of calcium supplementation on circulating lipids: potential

pharmacoeconomic implications. Drugs Aging, 2004. 21(1): p. 7-17.

20. Hu, F.B., J.E. Manson, and W.C. Willett, Types of Dietary Fat and Risk of Coronary

Heart Disease: A Critical Review. Journal of the American College of Nutrition, 2001.

20(1): p. 5-19.


association of calcium, phosphate, and bile acids. Gastroenterology, 1990. 99(6): p.

1653-9.

22. Major, G.C., et al., Supplementation with calcium + vitamin D enhances the beneficial

effect of weight loss on plasma lipid and lipoprotein concentrations. American Journal of

Clinical Nutrition, 2007. 85(1): p. 54-9.

179

23. Fantuzzi, G., Adipose tissue, adipokines, and inflammation. The Journal of allergy and

clinical immunology, 2005. 115(5): p. 911-9; quiz 920.

24. Van Lente, F., Markers of inflammation as predictors in cardiovascular disease. Clin

Chim Acta, 2000. 293(1-2): p. 31-52.

25. Plum, L.A. and H.F. DeLuca, Vitamin D, disease and therapeutic opportunities. Nat Rev

Drug Discov, 2010. 9(12): p. 941-55.

26. POLLY, P., et al., VDR-Alien: a novel, DNA-selective vitamin D3 receptor-corepressor

partnership. The FASEB Journal, 2000. 14(10): p. 1455-1463.

27. 2010 ICD-9-CM.


and its relation to BMI among Puerto Ricans Living in the Greater Boston Area. To be

submitted, 2013.


systemic inflammation in severe obese subjects. Internal and Emergency Medicine, 2013.

8(1): p. 33-40.


negatively correlated with serum 25(OH)D concentrations in healthy women. Journal of

Inflammation, 2008. 5(1): p. 10.


6, tumor necrosis factor-alpha and insulin resistance in overweight and obese subjects.

Cytokine, 2012. 60(3): p. 870-4.


suppress experimental inflammatory bowel disease. European Journal of Immunology,

2005. 35(1): p. 217-224.


cytokines production in PBMC modulating intracellular calcium. Transplantation

Proceedings, 2001. 33(3): p. 2366-2368.


macrophages from CAPD patients. Kidney International, 2001. 59(1): p. 69–75.

180

35. Alvarez-Rodriguez, L., et al., Age and low levels of circulating vitamin D are associated

with impaired innate immune function. Journal of Leukocyte Biology, 2012. 91(5): p.

829-838.

36. Dickie, L.J., et al., Vitamin D3 down-regulates intracellular Toll-like receptor 9

expression and Toll-like receptor 9-induced IL-6 production in human monocytes.

Rheumatology (Oxford), 2010. 49(8): p. 1466-1471.

37. Nonn, L., et al., Inhibition of p38 by Vitamin D Reduces Interleukin-6 Production in


Implications for Prostate Cancer Prevention by Vitamin D. Cancer Research, 2006.

66(8): p. 4516-4524.


with congestive heart failure: a double-blind, randomized, placebo-controlled trial.




Examination Survey 2001 to 2006). American Journal of Cardiology, 2012. 109(2): p.

226-30.

40. Gabay, C. and I. Kushner, Acute-phase proteins and other systemic responses to

inflammation. N Engl J Med, 1999. 340(6): p. 448-54.


Hispanic elders in Massachusetts. J Gerontol B Psychol Sci Soc Sci, 2000. 55(2): p.

S108-16.



43. Barcelo, A., et al., Waist circumference, BMI and the prevalence of self-reported diabetes

among the elderly of the United States and six cities of Latin America and the Caribbean.

Diabetes Res Clin Pract, 2007. 78(3): p. 418 - 427.

44. McTigue K, L.J.C.V.A. and et al., MOrtality and cardiac and vascular outcomes in

extremely obese women. JAMA, 2006. 296(1): p. 79-86.

181


levels: a population-based study in older men and women. Journal of Clinical

Endocrinology & Metabolism, 2005. 90(7): p. 4119-23.


markers of the metabolic syndrome and diabetes in overweight and obesity. Nutrition

Journal, 2008. 7: p. 4.

47. Young, K.A., et al., Association of Plasma Vitamin D Levels with Adiposity in Hispanic

and African Americans. Journal of Clinical Endocrinology & Metabolism, 2009. 94(9): p.

3306-3313.


concentrations: considerations and implications. Int J Obes, 2012. 36(3): p. 387- 96.

49. Wortsman, J., et al., Decreased bioavailability of vitamin D in obesity. The American

Journal of Clinical Nutrition, 2000. 72(3): p. 690-693.

50. Compher, C. and K.O. Badellino, Obesity and Inflammation: Lessons From Bariatric

Surgery. Journal of Parenteral and Enteral Nutrition, 2008. 32(6): p. 645-647.

51. Li, J., et al., 1α,25-Dihydroxyvitamin D hydroxylase in adipocytes. The Journal of Steroid

Biochemistry and Molecular Biology, 2008. 112(1–3): p. 122-126.


Endocrine Society Clinical Practice Guideline. Journal of Clinical Endocrinology &

Metabolism, 2011. 96(7): p. 1911-1930.

53. Looker, A.C.J., C L. Lacher, D A. Pfeiffer, C M. Schleicher, R L. Sempos, C T.,

Vitamin D status: United States, 2001–2006, in NCHS data brief2011, National Center

for Health Statistics: Hyattsville, MD.


in US adults. Nutrition Research, 2011. 31(1): p. 48-54.


56. Jamal-Allial, A.A., J.L. Griffith, and K.L. Tucker, The Longitudinal Association of

Vitamin D Serum Concentrations & Adiposity Phenotypes. To be submitted, 2013.

182

57. Renzaho, A.M.N., J.A. Halliday, and C. Nowson, Vitamin D, obesity, and obesity-related

chronic disease among ethnic minorities: A systematic review. Nutrition, 2011. 27(9): p.

868-879.


low vitamin D intake in a large US sample of healthy middle-aged men and women. J

Steroid Biochem Mol Biol, 2010. 121(1-2): p. 462-6.

59. Liel, Y., et al., Low circulating vitamin D in obesity. Calcified tissue international, 1988.

43(4): p. 199-201.

60. Lumb, G.A., E.B. Mawer, and S.W. Stanbury, The apparent vitamin D resistance of

chronic renal failure: A study of the physiology of vitamin D in man. The American

Journal of Medicine, 1971. 50(4): p. 421-441.

61. Mawer, E.B., et al., The distribution and storage of vitamin D and its metabolites in

human tissues. Clin Sci, 1972. 43(3): p. 413-31.

62. Jacobs, E.J., et al., WAist circumference and all-cause mortality in a large us cohort.

Archives of Internal Medicine, 2010. 170(15): p. 1293-1301.

63. Panagiotakos, D.B., et al., The implication of obesity and central fat on markers of

chronic inflammation: The ATTICA study. Atherosclerosis, 2005. 183(2): p. 308-315.

64. Dos Santos, R.E., et al., Relationship of body fat distribution by waist circumference,

dual-energy X-ray absorptiometry and ultrasonography to insulin resistance by

homeostasis model assessment and lipid profile in obese and non-obese postmenopausal

women. Gynecological Endocrinology, 2005. 21(5): p. 295-301.

65. Fontana, L., et al., Visceral Fat Adipokine Secretion Is Associated With Systemic

Inflammation in Obese Humans. Diabetes, 2007. 56(4): p. 1010-1013.

66. Junyent, M., et al., The effects of ABCG5/G8 polymorphisms on plasma HDL cholesterol

concentrations depend on smoking habit in the Boston Puerto Rican Health Study.

Journal of Lipid Research, 2009. 50(3): p. 565-73.

67. Lai, C.-Q., et al., Population admixture associated with disease prevalence in the Boston

Puerto Rican health study. Human Genetics, 2009. 125(2): p. 199-209.

183

68. Mattei, J., et al., Apolipoprotein A5 polymorphisms interact with total dietary fat intake in

association with markers of metabolic syndrome in Puerto Rican older adults. Journal of

Nutrition, 2009. 139(12): p. 2301-8.

69. Mattei, J., et al., Disparities in allele frequencies and population differentiation for 101

disease-associated single nucleotide polymorphisms between Puerto Ricans and non-

Hispanic whites. BMC Genetics, 2009. 14(10): p. 45.

70. Savgan-Gurol, E., et al., Waist to hip ratio and trunk to extremity fat (DXA) are better

surrogates for IMCL and for visceral fat respectively than for subcutaneous fat in

adolescent girls. Nutrition & Metabolism, 2010. 7(1): p. 86.

71. Holick, M.F., Vitamin D status: measurement, interpretation, and clinical application.

Ann Epidemiol, 2009. 19(2): p. 73-8.

72. Weishaar, T. and J.M. Vergili, Vitamin D Status Is a Biological Determinant of Health

Disparities. Journal of the Academy of Nutrition and Dietetics, 2013. 113(5): p. 643-651.


hydroxyvitamin D for multiple health outcomes. The American Journal of Clinical

Nutrition, 2006. 84(1): p. 18-28.

184

APPENDIX

185

Serum Vitamin D Concentration, Dietary Sources, and its relation to BMI among Puerto

Ricans Living in the Greater Boston Area1-2

Aziza Jamal-Allial3, Katherine L. Tucker 4*

3Pharmaceutical Science, Bouvé College of Health Sciences, Northeastern University. 4 Health

Sciences, Bouvé College of Health Sciences, Northeastern University, Boston, MA.

Word count: 447

FIGUREs: 4. Tables: 4. No OSM submitted

Running title: Vitamin D deficiency and obesity in Puerto Ricans

1 Supported by the National Institutes of Health (P01 AG023394, P50 HL105185, and R01

AG027087-01A1).

2 Author disclosures: A. Jamal-Allial, K. L. Tucker, no conflicts of interest.

Abbreviations used: Boston Puerto Rican Health Study (BPRHS), 25-hydroxyvitamin D

(25(OH)D), body mass index (BMI), National Health and Nutrition Examination Survey

(NHANES), Recommended Dietary Allowance (RDA).

*To whom correspondence should be addressed. Department of Health Sciences, Northeastern

University 208 Robinson Hall, 360 Huntington Avenue, Boston, MA 02115. Phone: (617) 373-

7952; Fax: (617) 373-2968. Email: [email protected]

186

ABSTRACT

Puerto Ricans have greater health disparities compared to the general population.

Vitamin D deficiency is highly prevalent among ethnic minorities. Cross-sectional analyses

included data from 1504 Puerto Ricans, 445 men and 1050 women, aged 45-75 years in Boston,

Massachusetts. Our aim was to determine serum 25-hydroxy vitamin D concentration (25(OH)D)

and associated contributing factors as well as test the hypothesis that lower serum 25(OH)D is

inversely associated with body mass index (BMI). Associations between 25(OH)D and likely

determinants were tested with multivariable linear regression. Low 25(OH)D concentration was

defined as <50 nmol/L. Dietary data from food frequency questionnaire were used to rank intake

sources of vitamin D using PROC RANK. Mean serum 25(OH)D concentration was 43.5 ± 16.5

nmol/L for men and 43.1 ± 16.0 nmol/L for women. 70.0% of participants had 25(OH)D <50

nmol/L. Milk products contributed 34% of total vitamin D intake, followed by supplements,

21.5%. Among women, but not men, higher intake of supplements was associated with

significantly higher serum 25(OH)D concentration. Never-smokers had significantly higher

serum 25(OH)D compared to current-smokers (P=0.01), but alcohol intake was not associated

with vitamin D status. The multivariate model showed that season (summer vs. others),

supplement use, dietary vitamin D intake, number of miles walked/day and serum albumin were

each positively associated (all P<0.05), while BMI was inversely associated (P=0.0007), with

serum 25(OH)D. Women with BMI ≥40 had significantly lower serum 25(OH)D concentration

compared to lower BMI. In conclusion, vitamin D deficiency is prevalent among Puerto Rican

adults living in Boston, MA, and this is associated with BMI and several behavioral measures,

including intake, smoking and outdoor walking.

187

INTRODUCTION

Vitamin D is naturally present in a limited number of food sources; but is available

through synthesis from its subcutaneous precursors with sun exposure[82]. Although 25-hydroxy

vitamin D [25(OH)D] is the measureable form and best clinical measure of status, there is no

consensus on a cutoff point for adequate vitamin D concentration. Most experts define deficiency

as 25(OH)D <50 nmol/L[115, 157]. Worldwide, vitamin D insufficiency is a major problem,

affecting more than 1 billion children and adults [26]. In the US, data from the Third National

Health and Nutrition Evaluation Survey (NHANES III) showed that approximately 41% of men

and 53% of women were considered vitamin D deficient [203]. A more recent NHANES report

indicated that vitamin D deficiency was more prevalent among blacks and Hispanics than non-

Hispanic whites, at 82.1% and 69.2% respectively [182]. A systematic review reported that the

mean of serum 25(OH)D for immigrants and/or ethnic minorities ranged between 24 to 60

nmol/L [171]. Vitamin D deficiency results in calcium and phosphorous abnormalities and may

lead to secondary hyperparathyroidism [115]. Due to its fat-soluble nature, adults with BMI >30

are at higher risk of vitamin D deficiency, which may be due, in part, to body fat sequestration

[115]. Several studies have reported associations between obesity and lower serum 25(OH)D

concentration [48, 49, 64, 173].

In the US, Puerto Ricans are the second largest Hispanic subgroup [204]. Recent

publications have confirmed the presence of important health disparities as well as low

socioeconomic status and education in this population [205]. Several reports have discussed the

high prevalence of obesity, type 2 diabetes, hypertension, depression and cognitive decline

among Puerto Ricans, relative to other Hispanic groups [164, 205-207]. However, we are aware

of no reports of vitamin D status in this population. Therefore, our aim was to describe the

188

distribution of serum 25(OH)D concentration among Puerto Rican men and women living in the

greater Boston area, and to explore its relationship with traditional determinants of 25(OH)D,

including BMI.


Study Population

Self-identified Puerto Rican adults, aged between 45-75 years, were recruited in the greater

Boston area as participants in the Boston Puerto Rican Health Study (BPRHS). The BPRHS is an

ongoing longitudinal study, started in 2005, to examine the role of social and psychological

stress on physiological dysfunction and related health outcomes among 1504 participants. Study

design, sampling, and recruitment methods have been described in detail [83]. Briefly, after

selecting census tracks that contained at least 25 Puerto Rican adults, aged 45 years and older,

blocks that contained ≥ 10 Hispanic adults, aged 45 and older, were enumerated. Other

recruitment methods were through random approach at community events (9.4%) or as

respondents to flyers or community members’ referral (6.9%). Participants were excluded if they

had serious health conditions or advanced dementia, had no permanent address, or planned to

move away within two years. The study protocol was approved by the Institutional Review

Boards at Tufts Medical Center and Northeastern University, and each participant provided

written informed consent.

For the current analysis, participants with missing data on variables needed for the analysis were

excluded (n=93) as well as those with serum 25(OH)D concentration greater than 250 nmol/L

(n=1) [55]. There were no significant differences in baseline sociodemographic characteristics

189

between those with complete or incomplete information. The present study includes 1400

participants with complete baseline data available at the time of analysis.

Field data collection

During home interviews, data on age, sex, education, household income, smoking status,

alcohol use, medical history and current medication use were collected by interviewer

administered questionnaire. Detailed description of the field data collection is available in

previous publications [83, 109]. Serum 25(OH)D concentration was categorized as deficient

when <50 nmol/L, insufficient between 50-75 nmol/L, and sufficient when > 75 nmol/L[26,

157]. BMI was calculated as weight (kg) divided by height (m)2 and defined by World Health

Organization criteria: recommended (BMI<25), overweight (25-30), obese (30-35), and very

obese (≥40) [208]. For subset analyses, we further categorized BMI using the class III obesity

definition [114]. Seasonality was defined as summer, July to September; fall, October to

December; winter, January to March; and spring, April to June. Age was categorized into two

groups: 45 to 59 years and 60-75 years. Both smoking status and alcohol were categorized into

yes and no for the current behavior.

Dietary vitamin D intake was estimated from a semi-quantitative food-frequency

questionnaire (FFQ) that was specifically designed for this population [91]. This FFQ has been

validated against serum carotenoids [116], and vitamin B12 [117] in a Hispanic population aged

≥ 60 years. Those with energy intakes < 600 or > 4800 kilocalories and/or > 10 questions blank

on the FFQ were excluded from the analyses. Nutrient intakes were calculated with the Nutrition

Data System for Research (NDS-R) software version 2007 (Nutrition Coordinating Center,

190

University of Minnesota, Minneapolis, MN, Food and Nutrient Database 2007, released 2007).

Vitamin D from supplements was defined as: minimal or no supplement < 5 µg/day (≤

200IU/day); moderate 5-10 µg/day (200-400 IU/day); and high ≥ 10 µg/day (≥ 400 IU/day).

Total vitamin D (food and supplements) was categorized according to the current RDA (15

µg/day or 600 IU/day) for vitamin D as: ≤ ½ the RDA (7.5 µg/day or 300 IU/day); >½ RDA- the

RDA (7.5-15 µg/day or 300-600 IU/day); and ≥ the RDA (10 µg/day or 600 IU/day) [209].

Dietary only vitamin D (µg/day) was categorized into quintiles. Serum creatinine concentration

(nmol/L) was categorized into quartiles. Serum albumin (g/L) was categorized as normal 34 – 54

g/L, or high, ≥ 54 g/L. Miles/day walked was calculated from blocks reported and categorized as:

0 miles/day, up to 0.5 mile/day, and more than 0.5 mile/day.

Laboratory assays



saved and stored at -80°C until processed. Serum 25(OH)D concentration were measured by

extraction followed by 25I radioimmunoassay Packard COBRA II Gamma Counter (DiaSorin

Inc., Stillwater, MN 55082) catalog # 68100E with intra-assay and inter-assay coefficients of

variation of 10.8% and 9.4%, respectively. Serum albumin was detected with a dye binding,

endpoint reaction using the Olympus AU400e with Olympus Albumin Reagents (OSR6102)

(Olympus America Inc., Melville, NY) Serum creatinine was assessed with a colorimetric,

kinetic reaction using the Olympus AU400e with Olympus Creatinine Reagents (OSR6178)

(Olympus America Inc., Melville, NY).

191

Statistical analyses

All statistical analyses were conducted with SAS 9.1.3 (SAS Institute Inc, Cary, NC); and

all tests were 2-sided with P <0.05 considered statistically significant. Skewed variables were

natural log-transformed to normalize their distributions before analysis. We calculated age-

adjusted mean serum 25(OH)D concentration across categories for known and potential

determinants of vitamin D status, and tested for linear trend with the general linear models

procedure (PROC GLM). Dunnett’s and Tukey’s tests were used for multiple comparison

adjustment. Linear regression analysis was used to quantify the association between age, sex,

education, poverty status, smoking status (yes/no), alcohol use (yes/no), number of miles

walked/day, BMI, serum albumin, and serum creatinine as independent variables, and serum

25(OH)D concentration as a dependent variable. Age-adjusted mean serum 25(OH)D was

calculated and plotted against medians for deciles of vitamin D intake. For comparison with

general U.S. adult dietary intakes, food groups were constructed according to the USDA

classification. Food sources of vitamin D intake were identified and ranked by the total intake

from foods. All dietary variables were energy adjusted

RESULTS

Complete serum 25(OH)D concentration and dietary intake information was available for

1316 participants, 935 women and 381 men. The mean of serum 25(OH)D among the BPRHS

participants was about 43.5± 16.1 nmol/L. Approximately 70% had low vitamin D status, <50

nmol/L, whereas only 4.2% of men and 3.5% of women had serum 25(OH)D concentration >75

nmol/L (Figure 6.1). The mean ages of men and women were 56.6 ± 8 and 57.3 ± 7.5,

192

respectively (Table 6.1). Women had significantly lower serum albumin, serum creatinine, and

walked fewer miles/day compared to men. In addition, women had significantly higher BMI and

poverty status than did men. More women than men were taking supplements with vitamin D,

and women tended to have higher total vitamin D intake. Serum 25(OH)D concentration was

associated with total vitamin D intake as defined by the RDA. The consumption of less than >½

RDA (7.5 µg/day) of total vitamin D intake was significantly lower than the >½ RDA –RDA and

≥ RDA increased serum 25(OH)D concentration significantly compared to lower amounts, 48

and 50 nmol/L compared to 40.3 nmol/L (Figure 6.2). The median intake of dietary vitamin D

(5.3 µg/day) and higher contributed to significantly higher serum 25(OH)D concentration

(nmol/L) compared to the referent group intake of 5.22 µg/day (P = 0.02, P = 0.005 and P =

0.01, respectively (Figure 6.3).

Men and women consuming daily supplements with at least 5µg/day of vitamin D had

significantly higher serum 25(OH)D concentration than those consuming supplements with less

than 5µg/day of vitamin D (P <0.0001) (Table 6.2). The season of examination was associated

with serum 25(OH)D concentration (P <0.0001), with lower concentration in the winter and

spring, relative to the summer. Higher serum creatinine concentration was associated with

significantly higher serum 25(OH)D concentration (P = 0.009 and P = 0.02 for men and women,

respectively). Among women only, older women, aged 60-75 years, had significantly higher

serum 25(OH)D than did women aged 45-59 y (P = 0.007). BMI had an inverse relationship with

age-adjusted serum 25(OH)D concentration (P-trend (0.001). Compared to never-smokers,

women who were current smokers had significantly lower 25(OH)D concentration (P = 0.02).

Women, but not men, had significantly higher serum 25(OH)D concentration when walking

more miles/day (P-trend =0.01). Among men, lower serum albumin concentration was associated

193

with lower serum 25(OH)D (P= 0.05). Alcohol intake was not significantly associated with

25(OH)D.

In a full model with adjustment for other variables, serum 25(OH)D remained

significantly associated with dietary vitamin D intake (P <0.0001), use of vitamin D-containing

supplements (P <0.0001), seasonality (P <0.0001), and number of miles walked/day (P <0.05)

(Table 6.3). There was a significant interaction between sex and vitamin D supplement use (P =

0.008) (Figure 6.4). Among women, supplements with vitamin D in the range of 4.5-10 µg/day

and those ≥ 10 µg/day, were associated with significantly higher mean serum 25(OH)D

concentration, 45 nmol and 51.3 nmol/L, respectively, compared with non-supplement use (P

<0.0001), 40 nmol/L. However, among men, only supplements providing 5-10 µg/day of vitamin

D were associated with significantly higher mean serum 25(OH)D concentration compared with

non-supplement use (P <0.0001).

Another interaction was between seasonality and age (P = 0.0008) (Figure 6.5). Younger

participants, ages 45-59 years, had significantly higher 25(OH)D in both summer and fall, 50.5

nmol/L and 49.0 nmol/L, respectively, relative to winter and spring, (41.5 nmol/L and 39.0

nmol/L, respectively, P <0.0001). However, among older participants, ages 56 - 75 years, serum

25(OH)D did not differ significantly across the four seasons.

Subset analyses of the women showed that serum 25(OH)D was significantly lower with

increased obesity (P-trend = 0.001) when BMI was further categorized into obesity categories.

Women with class III obesity, BMI ≥ 40, had significantly lower serum 25(OH)D concentration

(41.6 nmol/L) compared to lower BMI (Figure 6.4) adjusting for age, seasonality and vitamin D

supplements, and number of miles walked/day. In addition to supplements, a range of foods

contributed to vitamin D intake (Table 6.4). Fortified milk was the major contributor, at 34% of

194

total vitamin D intake from food. Fish ranked second, at 16.2%, as one of the food sources that is

naturally rich of vitamin D. Eggs and ready-to-eat cereal each contributed about 4.5% to total

dietary vitamin D intake.

DISCUSSION

These results highlight the high prevalence of vitamin D deficiency among Puerto Rican

adults living in the greater Boston area, and show an inverse relationship between BMI and

serum 25(OH)D concentration. The majority (70%) of the men and women in our study had low

vitamin D status, with a mean circulating of 25(OH)D of 44.0 nmol/L—which is lower than that

reported in the literature for several other populations [182, 210-212]. Our data concur with a

recent NHANES report showing that 69.2% of U.S. Hispanics had low vitamin D. This greater

risk of vitamin D deficiency may translate into greater risk of health disparities [182]. Although

debated, most current recommendations suggest that serum concentration should be above 75

nmol/L [26, 29, 51, 82, 182]. We found that only 3% of Puerto Rican men and 4% of women

achieved this [213]. This disparity may have various explanations, including skin pigmentation

[31, 182]. Darker skin is associated with lower 7-dehydrocholesterol conversion to its active

metabolite, which requires longer sunlight exposure [214]. In addition, these participants reside

in Boston, which lies on the 43.3°N, with relatively low UV-B exposure [31]. Moreover, 50% of

the BPRHS participants reported incomes below the poverty line, and 60% reported education of

less than 8th grade [83] indicating low socioeconomic status. This may translate into fewer

healthy food choices and limited access to supplements.

195

Unlike many reports, aging was not negatively associated with serum 25(OH)D

concentration in this population [212, 215] . With aging, the presence of vitamin D precursors

within the skin declines, which puts the elderly at higher risk of vitamin D deficiency [26, 216].

A recent report using NHANES III data reported significant declines with aging [215, 217] while

others have reported no significant difference across age groups [182]. In our sample, the

difference appears to be due to greater use of supplements, particularly among older women,

relative to younger individuals. Vitamin D supplements as well as dietary vitamin D have

previously been shown to be associated with serum 25(OH)D concentration[218, 219]. This

study confirms the importance of vitamin D from supplementation and dietary sources,

especially among women, on improving vitamin D status. In our population, 30% of women

were consuming supplements with higher vitamin D concentration (≥ 5 µg/day or 200 IU/day)

compared to 24% of men. In general, women tend to be more likely than men to be prescribed

vitamin D-containing supplements, often for osteoporosis prevention. When compared to the

RDA, levels, only 5% of this population is meeting the current RDA of vitamin D, which is 15

µg/day or 600 IU/day [209]. This low percentage meeting the RDA is consistent with poor health

status, which has been observed among this population [109, 113, 220].

The relationship between BMI and serum 25(OH)D concentration among Puerto Rican

women is noteworthy. Our results showed that with increased obesity, especially when reaching

class III obesity, serum 25(OH)D concentration declined significantly. Our results indicate that

serum 25(OH)D concentration reached a plateau at BMI 35 or less, compared to severely obese

categories. Most studies have found an inverse relationship between obesity and serum 25(OH)D

concentration [48, 64, 173]. One of the proposed theories that links obesity to lower 25(OH)D is

that the latter leads to an increase in iPTH concentration, resulting in interference with lipolysis

196

and greater expression of fatty acid synthase [221]. Another theory suggests that lower 25(OH)D

may be one of the causes of increased fat mass during the winter season [222]. It has also been

shown that there is increased sequestration of 7-dehydrocholesterol in adipose tissue, thus fewer

precursor molecules are available [48, 49]. This population has a high (57%) prevalence of

obesity [83, 113, 223], combined with very low physical activity [83], both of which translate

into low outdoor activity, thus reduced sun exposure, all of which are strong predictors of lower

serum 25(OH)D concentration. This complex of factors may explain why this Puerto Rican

population has very low serum 25(OH)D concentration.

Several limitations of our study should be noted. First, due to the cross-sectional design,

associations cannot be translated into cause and effect relationships. In addition, residual

confounding remains a possibility, even though we adjusted for several essential covariates.

Third, we did not have a measure of sun exposure, which would have added predictive ability to

the models, although we did include a measure of outdoor physical activity which is a likely

proxy. Nevertheless, our study design was strengthened by the use of our validated FFQ that has

been designed for this particular ethnicity group [116, 117]. This report of vitamin D status

shows high prevalence of vitamin D deficiency in a large Puerto Rican adult population. Because

vitamin D status has been associated with numerous health outcomes, it is likely that improving

vitamin D status in this population at high risk of health disparities, could contribute to reducing

health disparities.

197

BIBLIOGRAPHY

1. Organization, W.H., Cardiovascular diseases (CVDs), in Fact sheet N°3172011, WHO: Geneva.

2. Balagopal, P., et al., Nontraditional Risk Factors and Biomarkers for Cardiovascular

Disease: Mechanistic, Research, and Clinical Considerations for Youth. Circulation, 2011. 123(23): p. 2749-2769.

3. Lusis, A.J., Atherosclerosis. Nature, 2000. 407(6801): p. 233-241. 4. Krishnan, A.V. and D. Feldman, Mechanisms of the anti-cancer and anti-inflammatory

actions of vitamin D. Annual review of pharmacology and toxicology, 2011. 51: p. 311-36.

5. Tomaschitz, A., et al., Independent association between 1,25-dihydroxyvitamin D, 25-

hydroxyvitamin D and the renin–angiotensin system: The Ludwigshafen Risk and

Cardiovascular Health (LURIC) study. Clinica Chimica Acta, 2010. 411(17–18): p. 1354-1360.


Cardiovascular Health. The American Journal of Cardiology, 2012. 109(3): p. 359-363. 7. Sherry L. Murphy, B.S.J.X., M.D.; and Kenneth D. Kochanek, M.A., Deaths:

Preliminary Data for 2010, in National Vital Statisictis Report2012, CDC: Division of Vital Statistics.

8. Writing Group, M., et al., Heart disease and stroke statistics--2010 update: a report from

the American Heart Association. Circulation, 2010. 121(7): p. e46-e215. 9. Members, W.G., et al., Executive Summary: Heart Disease and Stroke Statistics—2011

Update. Circulation, 2011. 123(4): p. 459-463. 10. Wilson, P.W. and B.F. Culleton, Epidemiology of cardiovascular disease in the United

States. American Journal of Kidney Diseases, 1998. 32(5 Suppl 3). 11. Lloyd-Jones, D.M., et al., Lifetime risk for developing congestive heart failure: the

Framingham Heart Study. Circulation, 2002. 106(24): p. 3068-72. 12. Tedgui, A. and Z. Mallat, Cytokines in Atherosclerosis: Pathogenic and Regulatory

Pathways. Physiological Reviews, 2006. 86(2): p. 515-581. 13. Ross, R., Atherosclerosis is an inflammatory disease. Am Heart J, 1999. 138: p. S419-20. 14. Davignon, J. and P. Ganz, Role of Endothelial Dysfunction in Atherosclerosis.

Circulation, 2004. 109(23 suppl 1): p. III-27-III-32. 15. Lind, L., Lipids and endothelium-dependent vasodilation—A review. Lipids, 2002. 37(1):

p. 1-15. 16. Deshpande, D.D., D.R. Janero, and M.M. Amiji, Therapeutic strategies for endothelial

dysfunction. Expert opinion on biological therapy, 2011. 11(12): p. 1637-54. 17. Ridker, P.M., Clinical application of C-reactive protein for cardiovascular disease

detection and prevention. Circulation, 2003. 107(3): p. 363-9. 18. Shishehbor, M.H. and D.L. Bhatt, Inflammation and atherosclerosis. Current

Atherosclerosis Reports, 2004. 6(2): p. 131-9. 19. Almirall, J., et al., Association of low serum 25-hydroxyvitamin D levels and high arterial

blood pressure in the elderly. Nephrology Dialysis Transplantation, 2010. 25(2): p. 503-9.

198

20. Ameri, P., et al., High prevalence of vitamin D deficiency and its association with left

ventricular dilation: an echocardiography study in elderly patients with chronic heart

failure. Nutrition Metabolism & Cardiovascular Diseases, 2010. 20(9): p. 633-40. 21. Baz-Hecht, M. and A.B. Goldfine, The impact of vitamin D deficiency on diabetes and

cardiovascular risk. Current Opinion in Endocrinology, Diabetes & Obesity, 2010. 17(2): p. 113-9.

22. Bostick, R.M., et al., Relation of calcium, vitamin D, and dairy food intake to ischemic

heart disease mortality among postmenopausal women. American Journal of Epidemiology, 1999. 149(2): p. 151-61.

23. Giovannucci, E., et al., 25-hydroxyvitamin D and risk of myocardial infarction in men: a

prospective study. Archives of Internal Medicine, 2008. 168(11): p. 1174-80. 24. Oh, J., et al., 1,25(OH)2 Vitamin D Inhibits Foam Cell Formation and Suppresses

Macrophage Cholesterol Uptake in Patients With Type 2 Diabetes Mellitus. Circulation, 2009. 120(8): p. 687-698.

25. Sugden, J.A., et al., Vitamin D improves endothelial function in patients with Type 2

diabetes mellitus and low vitamin D levels. Diabet Med, 2008. 25(3): p. 320-5. 26. Holick, M.F., Vitamin D deficiency. N Engl J Med, 2007. 357(3): p. 266-81. 27. Armas, L.A., B.W. Hollis, and R.P. Heaney, Vitamin D2 is much less effective than

vitamin D3 in humans. Journal of Clinical Endocrinology & Metabolism, 2004. 89(11): p. 5387-91.

28. Holick, M.F., et al., The vitamin D content of fortified milk and infant formula. N Engl J Med, 1992. 326(18): p. 1178-81.

29. Holick, M.F., High prevalence of vitamin D inadequacy and implications for health. Mayo Clin Proc, 2006. 81(3): p. 353-73.

30. Holick, M.F., Resurrection of vitamin D deficiency and rickets. J Clin Invest, 2006. 116(8): p. 2062-72.

31. Holick, M.F., Deficiency of sunlight and vitamin D. Bmj, 2008. 336(7657): p. 1318-9. 32. Ramos-Lopez, E., et al., CYP2R1 (vitamin D 25-hydroxylase) gene is associated with

susceptibility to type 1 diabetes and vitamin D levels in Germans. Diabetes/Metabolism Research and Reviews, 2007. 23(8): p. 631-636.

33. Rhieu, S.Y., et al., A new insight into the role of rat cytochrome P450 24A1 in

metabolism of selective analogs of 1alpha,25-dihydroxyvitamin D. Archives of biochemistry and biophysics, 2011. 509(1): p. 33-43.


vitamin D. Trends in biochemical sciences, 2004. 29(12): p. 664-73. 35. Cheng, J.B., et al., Genetic evidence that the human CYP2R1 enzyme is a key vitamin D

25-hydroxylase. Proceedings of the National Academy of Sciences of the United States of America, 2004. 101(20): p. 7711-5.

36. Jones, G., Pharmacokinetics of vitamin D toxicity. The American Journal of Clinical Nutrition, 2008. 88(2): p. 582S-586S.

37. Deeb, K.K., D.L. Trump, and C.S. Johnson, Vitamin D signalling pathways in cancer:

potential for anticancer therapeutics. Nature reviews. Cancer, 2007. 7(9): p. 684-700. 38. Plum, L.A. and H.F. DeLuca, Vitamin D, disease and therapeutic opportunities. Nat Rev

Drug Discov, 2010. 9(12): p. 941-55.

199

39. POLLY, P., et al., VDR-Alien: a novel, DNA-selective vitamin D3 receptor-corepressor

partnership. The FASEB Journal, 2000. 14(10): p. 1455-1463. 40. Holick, M.F., Evolution and function of vitamin D. Recent Results Cancer Res, 2003.

164: p. 3-28. 41. Holick, M.F., The influence of vitamin D on bone health across the life cycle. J Nutr,

2005. 135(11): p. 2726S-7S. 42. Etten, E.v. and C. Mathieu, Immunoregulation by 1,25-dihydroxyvitamin D3: Basic

concepts. The Journal of Steroid Biochemistry and Molecular Biology, 2005. 97(1–2): p. 93-101.

43. Bertone-Johnson, E.R., et al., Plasma 25-hydroxyvitamin D and 1,25-dihydroxyvitamin D

and risk of breast cancer. Cancer Epidemiol Biomarkers Prev, 2005. 14(8): p. 1991-7. 44. Chen, T.C., et al., Prostatic 25-hydroxyvitamin D-1alpha-hydroxylase and its implication

in prostate cancer. J Cell Biochem, 2003. 88(2): p. 315-22. 45. Chen, T.C. and M.F. Holick, Vitamin D and prostate cancer prevention and treatment.

Trends Endocrinol Metab, 2003. 14(9): p. 423-30. 46. Judd, S.E. and V. Tangpricha, Vitamin D deficiency and risk for cardiovascular disease.

American Journal of the Medical Sciences, 2009. 338(1): p. 40-4. 47. Kim, D.H., et al., Prevalence of hypovitaminosis D in cardiovascular diseases (from the

National Health and Nutrition Examination Survey 2001 to 2004). American Journal of Cardiology, 2008. 102(11): p. 1540-4.

48. Wortsman, J., et al., Decreased bioavailability of vitamin D in obesity. The American Journal of Clinical Nutrition, 2000. 72(3): p. 690-693.

49. Liel, Y., et al., Low circulating vitamin D in obesity. Calcified tissue international, 1988. 43(4): p. 199-201.

50. Michael F, H.P.D., M.D., ed. Vitamin D. Physiology, Molecular Biology, and clinical

applications. Second ed. NUTRITION AND HEALTH, ed. P. Adrianne Bendich. 2010, Springer Humana Press. 1155.


hydroxyvitamin D for multiple health outcomes. American Journal of Clinical Nutrition, 2006. 84(1): p. 18-28.

52. Chapuy, M.C., et al., Prevalence of vitamin D insufficiency in an adult normal

population. Osteoporos Int, 1997. 7(5): p. 439-43. 53. Malabanan, A., I.E. Veronikis, and M.F. Holick, Redefining vitamin D insufficiency.

Lancet, 1998. 351(9105): p. 805-6. 54. Dawson-Hughes, B., et al., Estimates of optimal vitamin D status. Osteoporos Int, 2005.

16(7): p. 713-6. 55. Holick, M.F., Vitamin D status: measurement, interpretation, and clinical application.

Ann Epidemiol, 2009. 19(2): p. 73-8. 56. Holick, M.F. and T.C. Chen, Vitamin D deficiency: a worldwide problem with health

consequences. Am J Clin Nutr, 2008. 87(4): p. 1080S-6S. 57. Gordon, C.M., et al., Prevalence of vitamin D deficiency among healthy adolescents.

Arch Pediatr Adolesc Med, 2004. 158(6): p. 531-7. 58. Sullivan, S.S., et al., Adolescent girls in Maine are at risk for vitamin D insufficiency. J

Am Diet Assoc, 2005. 105(6): p. 971-4.

200

59. Araujo, A., et al., Serum 25-hydroxyvitamin D and bone mineral density among Hispanic

men. Osteoporosis International, 2009. 20(2): p. 245-255. 60. Palacios, C., et al., Determinants of vitamin D status among overweight and obese Puerto

Rican adults. Ann Nutr Metab, 2012. 60(1): p. 35-43. 61. Zgaga, L., et al., Diet, environmental factors, and lifestyle underlie the high prevalence of

vitamin D deficiency in healthy adults in Scotland, and supplementation reduces the

proportion that are severely deficient. Journal of Nutrition, 2011. 141(8): p. 1535-42. 62. Michos, E.D., et al., 25-Hydroxyvitamin D deficiency is associated with fatal stroke

among whites but not blacks: The NHANES-III linked mortality files. Nutrition, (0). 63. Freedman, D.M., et al., Serum Levels of Vitamin D Metabolites and Breast Cancer Risk

in the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial. Cancer Epidemiology Biomarkers & Prevention, 2008. 17(4): p. 889-894.

64. Botella-Carretero, J.I., et al., Vitamin D deficiency is associated with the metabolic

syndrome in morbid obesity. Clinical Nutrition, 2007. 26(5): p. 573-80. 65. Oh, J.H., et al., The level of vitamin D in the serum correlates with fatty degeneration of

the muscles of the rotator cuff. Journal of Bone & Joint Surgery - British Volume, 2009. 91(12): p. 1587-93.

66. Visser, M., et al., Low vitamin D and high parathyroid hormone levels as determinants of

loss of muscle strength and muscle mass (sarcopenia): the Longitudinal Aging Study

Amsterdam. Journal of Clinical Endocrinology & Metabolism, 2003. 88(12): p. 5766-72. 67. Kragt, J., et al., Higher levels of 25-hydroxyvitamin D are associated with a lower

incidence of multiple sclerosis only in women. Multiple Sclerosis, 2009. 15(1): p. 9-15. 68. Wang, T.J., et al., Vitamin D deficiency and risk of cardiovascular disease. Circulation,

2008. 117(4): p. 503-11. 69. Pilz, S., et al., Association of Vitamin D Deficiency with Heart Failure and Sudden

Cardiac Death in a Large Cross-Sectional Study of Patients Referred for Coronary

Angiography. Journal of Clinical Endocrinology & Metabolism, 2008. 93(10): p. 3927-3935.

70. Autier, P. and S. Gandini, Vitamin D supplementation and total mortality: a meta-

analysis of randomized controlled trials. Archives of Internal Medicine, 2007. 167(16): p. 1730-7.

71. Chung, I., et al., Anti-proliferative effects of calcitriol on endothelial cells derived from

two different microenvironments. The Journal of Steroid Biochemistry and Molecular Biology, 2007. 103(3–5): p. 768-770.

72. Jacques M, L., Immunomodulatory actions of 1,25-Dihydroxyvitamin D3. The Journal of Steroid Biochemistry and Molecular Biology, 1995. 53(1–6): p. 599-602.

73. Berk, B.C., et al., Endothelial Atheroprotective and Anti-inflammatory Mechanisms. Annals of the New York Academy of Sciences, 2001. 947(1): p. 93-111.

74. de Vries-van der Weij, J., et al., Anti-inflammatory salicylate beneficially modulates pre-

existing atherosclerosis through quenching of NF-κB activity and lowering of

cholesterol. Atherosclerosis, 2010. 213(1): p. 241-246. 75. Talmor, Y., et al., Calcitriol blunts pro-atherosclerotic parameters through NFkappaB

and p38 in vitro. Eur J Clin Invest, 2008. 38(8): p. 548-54.

201

76. Lumb, G.A., E.B. Mawer, and S.W. Stanbury, The apparent vitamin D resistance of

chronic renal failure: A study of the physiology of vitamin D in man. The American Journal of Medicine, 1971. 50(4): p. 421-441.

77. Mawer, E.B., et al., The distribution and storage of vitamin D and its metabolites in

human tissues. Clin Sci, 1972. 43(3): p. 413-31. 78. Becker, L.E., et al., Effect of paricalcitol and calcitriol on aortic wall remodeling in

uninephrectomized ApoE knockout mice. American Journal of Physiology - Renal Physiology, 2011. 300(3): p. F772-F782.

79. Husain, K., et al., Effects of Paricalcitol and Enalapril on Atherosclerotic Injury in

Mouse Aortas. American Journal of Nephrology, 2010. 32(4): p. 296-304. 80. Takeda, M., et al., Oral Administration of an Active Form of Vitamin D3 (Calcitriol)

Decreases Atherosclerosis in Mice by Inducing Regulatory T Cells and Immature

Dendritic Cells With Tolerogenic Functions. Arteriosclerosis, Thrombosis, and Vascular Biology, 2010. 30(12): p. 2495-2503.

81. Wang, X.X., et al., Vitamin D receptor agonist doxercalciferol modulates dietary fat-

induced renal disease and renal lipid metabolism. American Journal of Physiology - Renal Physiology, 2011. 300(3): p. F801-F810.

82. Dietary Reference Intakes for Calcium and Vitamin D. Committee to Review Dietary

Reference Intakes for Vitamin D and Calcium. , in 2010, Institute of Medicine: Washington, DC: National Academy of Sciences.

83. Tucker, K., et al., The Boston Puerto Rican Health Study, a longitudinal cohort study on

health disparities in Puerto Rican adults: challenges and opportunities. BMC Public Health, 2010. 10(1): p. 107.

84. Bermudez, O.I., E.K. Becker, and K.L. Tucker, Development of sex-specific equations for

estimating stature of frail elderly Hispanics living in the northeastern United States. American Journal of Clinical Nutrition, 1999. 69(5): p. 992-8.

85. Chumlea, W., et al., Stature prediction equations for elderly non-Hispanic white, non-

Hispanic black, and Mexican-American persons developed from NHANES III data. J Am Diet Assoc, 1998. 98: p. 137 - 142.

86. Allain, C.C., et al., Enzymatic determination of total serum cholesterol. Clin Chem, 1974. 20(4): p. 470-5.

87. Friedewald, W.T., R.I. Levy, and D.S. Fredrickson, Estimation of the concentration of

low-density lipoprotein cholesterol in plasma, without use of the preparative

ultracentrifuge. Clin Chem, 1972. 18(6): p. 499-502. 88. Dreon, D., et al., Use of NHANES Data to Assign Nutrient Densities to Food Groups in a

Multiethnic Diet History Questionnaire. Nutrition and Cancer, 1993. 20: p. 223 - 230. 89. Delgado, J.L., et al., Hispanic Health and Nutrition Examination Survey: methodological

considerations. Am J Public Health, 1990. 80 Suppl: p. 6-10. 90. Block, G. and A.F. Subar, Estimates of nutrient intake from a food frequency

questionnaire: the 1987 National Health Interview Survey. J Am Diet Assoc, 1992. 92(8): p. 969-77.



202

92. Brecher, G., M. Schneiderman, and G.Z. Williams, Evaluation of electronic red blood

cell counter. Am J Clin Pathol, 1956. 26(12): p. 1439-49. 93. Holick, M.F., The vitamin D deficiency pandemic and consequences for nonskeletal

health: mechanisms of action. Mol Aspects Med, 2008. 29(6): p. 361-8. 94. Holick, M.F., Vitamin D: extraskeletal health. Endocrinology and metabolism clinics of

North America, 2010. 39(2): p. 381-400, table of contents. 95. Lacour, B., et al., Parathyroid function and lipid metabolism in the rat. Miner Electrolyte

Metab, 1982. 7(3): p. 157-65. 96. Melamed, M.L., et al., 25-hydroxyvitamin d levels and the risk of mortality in the general

population. Archives of Internal Medicine, 2008. 168(15): p. 1629-1637. 97. Alvarez-Rodriguez, L., et al., Age and low levels of circulating vitamin D are associated

with impaired innate immune function. Journal of Leukocyte Biology, 2012. 91(5): p. 829-838.



99. Pittas, A.G., et al., Systematic Review: Vitamin D and Cardiometabolic Outcomes. Annals of Internal Medicine, 2010. 152(5): p. 307-314.


weight loss on cardiovascular disease risk markers. American Journal of Clinical Nutrition, 2009. 89(5): p. 1321-7.

101. Kannel, W.B., et al., Serum Cholesterol, Lipoproteins, and the Risk of Coronary Heart

DiseaseThe Framingham Study. Annals of Internal Medicine, 1971. 74(1): p. 1-12. 102. Gordon, T., et al., High density lipoprotein as a protective factor against coronary heart

disease: The Framingham study. The American Journal of Medicine, 1977. 62(5): p. 707-714.

103. Hypponen, E., et al., 25-hydroxyvitamin D, IGF-1, and metabolic syndrome at 45 years

of age: a cross-sectional study in the 1958 British Birth Cohort. Diabetes, 2008. 57(2): p. 298-305.

104. Jorde, R., et al., High serum 25-hydroxyvitamin D concentrations are associated with a

favorable serum lipid profile. European Journal of Clinical Nutrition, 2010. 64(12): p. 1457-64.

105. Karhapää, P., et al., Diverse associations of 25-hydroxyvitamin D and 1,25-dihydroxy-

vitamin D with dyslipidaemias. Journal of Internal Medicine, 2010. 268(6): p. 604-610. 106. Kazlauskaite, R., et al., Vitamin D is associated with atheroprotective high-density

lipoprotein profile in postmenopausal women. Journal of Clinical Lipidology, 2010. 4(2): p. 113-119.

107. Maki, K.C., et al., Serum 25-hydroxyvitamin D is independently associated with high-

density lipoprotein cholesterol and the metabolic syndrome in men and women. Journal of Clinical Lipidology, 2009. 3(4): p. 289-296.

108. Jorde, R., et al., No improvement in cardiovascular risk factors in overweight and obese

subjects after supplementation with vitamin D3 for 1 year. Journal of Internal Medicine, 2010. 267(5): p. 462-472.



203



white neighborhood comparison group. J Gerontol A Biol Sci Med Sci, 2000. 55(2): p. M90-7.


controlled among Hispanic elders of Caribbean origin. Am J Public Health, 2000. 90(8): p. 1288-93.

112. Renzaho, A.M., J.A. Halliday, and C. Nowson, Vitamin D, obesity, and obesity-related




114. Flegal, K.M., et al., Overweight and obesity in the United States: prevalence and trends,

1960-1994. International journal of obesity and related metabolic disorders : journal of the International Association for the Study of Obesity, 1998. 22(1): p. 39-47.


Endocrine Society Clinical Practice Guideline. Journal of Clinical Endocrinology & Metabolism, 2011. 96(7): p. 1911-1930.


different patterns of carotenoid intake and plasma concentrations. Journal of Nutrition, 2005. 135(6): p. 1496 - 1502.


more prevalent in Hispanic older adults of Caribbean origin than in neighborhood-

matched non-Hispanic whites. Journal of Nutrition, 2002. 132(7): p. 2059-64. 118. Pharmacists, A.S.o.H.-S., AHFS Pharmacologic-Therapeutic Classification 2008,

American Society of Health-System Pharmacists. 119. Friedewald, W.T., R.I. Levy, and D.S. Fredrickson, Estimation of the Concentration of

Low-Density Lipoprotein Cholesterol in Plasma, Without Use of the Preparative

Ultracentrifuge. Clinical Chemistry, 1972. 18(6): p. 499-502. 120. Signorello, L.B., et al., Confounding by indication in epidemiologic studies of commonly

used analgesics. American Journal of Therapeutics, 2002. 9(3): p. 199-205. 121. Barnett, A.G., J.C. van der Pols, and A.J. Dobson, Regression to the mean: what it is and

how to deal with it. International Journal of Epidemiology, 2005. 34(1): p. 215-220. 122. Sterne, J.A.C., et al., Multiple imputation for missing data in epidemiological and clinical

research: potential and pitfalls. BMJ, 2009. 338. 123. Delvin, E.E., et al., Vitamin D Status Is Modestly Associated with Glycemia and

Indicators of Lipid Metabolism in French-Canadian Children and Adolescents. The Journal of Nutrition, 2010. 140(5): p. 987-991.

124. Forouhi, N.G., et al., Baseline Serum 25-Hydroxy Vitamin D Is Predictive of Future

Glycemic Status and Insulin Resistance: The Medical Research Council Ely Prospective

Study 1990–2000. Diabetes, 2008. 57(10): p. 2619-2625. 125. Giovannucci, E., et al., 25-hydroxyvitamin d and risk of myocardial infarction in men: A

prospective study. Archives of Internal Medicine, 2008. 168(11): p. 1174-1180.

204

126. Pilz, S., et al., Vitamin D and mortality in older men and women. Clinical Endocrinology, 2009. 71(5): p. 666-672.

127. Williams, D.M., A. Fraser, and D.A. Lawlor, Associations of vitamin D, parathyroid

hormone and calcium with cardiovascular risk factors in US adolescents. Heart, 2011. 97(4): p. 315-320.

128. Bonow, R.O., Primary Prevention of Cardiovascular Disease. Circulation, 2002. 106(25): p. 3140-3141.


weight loss on cardiovascular disease risk markers. The American Journal of Clinical Nutrition, 2009. 89(5): p. 1321-1327.

130. Heikkinen, A., et al., Long-term vitamin D3 supplementation may have adverse effects on

serum lipids during postmenopausal hormone replacement therapy. European Journal of Endocrinology, 1997. 137(5): p. 495-502.

131. Nagpal, J., J.N. Pande, and A. Bhartia, A double-blind, randomized, placebo-controlled

trial of the short-term effect of vitamin D3 supplementation on insulin sensitivity in

apparently healthy, middle-aged, centrally obese men. Diabetic Medicine, 2009. 26(1): p. 19-27.


hydroxyvitamin D and body mass index: the Tromsø study. European Journal of Nutrition, 2010. 49(7): p. 401-407.

133. Yoon, V., N.M. Maalouf, and K. Sakhaee, The effects of smoking on bone metabolism. Osteoporosis International, 2012. 23(8): p. 2081-2092.

134. Lakhan, S. and A. Kirchgessner, Anti-inflammatory effects of nicotine in obesity and

ulcerative colitis. Journal of Translational Medicine, 2011. 9(1): p. 129. 135. Wang, H., et al., Nicotinic acetylcholine receptor alpha7 subunit is an essential regulator

of inflammation. Nature, 2003. 421(6921): p. 384-8. 136. Reid, I.R., Effects of calcium supplementation on circulating lipids: potential

pharmacoeconomic implications. Drugs Aging, 2004. 21(1): p. 7-17. 137. Hu, F.B., J.E. Manson, and W.C. Willett, Types of Dietary Fat and Risk of Coronary

Heart Disease: A Critical Review. Journal of the American College of Nutrition, 2001. 20(1): p. 5-19.


association of calcium, phosphate, and bile acids. Gastroenterology, 1990. 99(6): p. 1653-9.

139. Major, G.C., et al., Supplementation with calcium + vitamin D enhances the beneficial

effect of weight loss on plasma lipid and lipoprotein concentrations. American Journal of Clinical Nutrition, 2007. 85(1): p. 54-9.

140. Junyent, M., et al., The effects of ABCG5/G8 polymorphisms on plasma HDL cholesterol

concentrations depend on smoking habit in the Boston Puerto Rican Health Study. Journal of Lipid Research, 2009. 50(3): p. 565-73.

141. Lai, C.-Q., et al., Population admixture associated with disease prevalence in the Boston

Puerto Rican health study. Human Genetics, 2009. 125(2): p. 199-209. 142. Mattei, J., et al., Apolipoprotein A5 polymorphisms interact with total dietary fat intake in

association with markers of metabolic syndrome in Puerto Rican older adults. Journal of Nutrition, 2009. 139(12): p. 2301-8.

205

143. Mattei, J., et al., Disparities in allele frequencies and population differentiation for 101

disease-associated single nucleotide polymorphisms between Puerto Ricans and non-

Hispanic whites. BMC Genetics, 2009. 14(10): p. 45. 144. 2010 ICD-9-CM.

145. Guillot, X., et al., Vitamin D and inflammation. Joint Bone Spine, 2010. 77(6): p. 552-557.

146. Querfeld, U., Vitamin D and inflammation. Pediatr Nephrol, 2013. 28(4): p. 605-10. 147. Dickie, L.J., et al., Vitamin D3 down-regulates intracellular Toll-like receptor 9

expression and Toll-like receptor 9-induced IL-6 production in human monocytes. Rheumatology (Oxford), 2010. 49(8): p. 1466-1471.

148. Wilson, A.M., M.C. Ryan, and A.J. Boyle, The novel role of C-reactive protein in

cardiovascular disease: risk marker or pathogen. Int J Cardiol, 2006. 106(3): p. 291-7. 149. Gabay, C. and I. Kushner, Acute-phase proteins and other systemic responses to

inflammation. N Engl J Med, 1999. 340(6): p. 448-54. 150. Rigby, W.F., S. Denome, and M.W. Fanger, Regulation of lymphokine production and

human T lymphocyte activation by 1,25-dihydroxyvitamin D3. Specific inhibition at the

level of messenger RNA. The Journal of Clinical Investigation, 1987. 79(6): p. 1659-1664.

151. Rigby, W.F., T. Stacy, and M.W. Fanger, Inhibition of T lymphocyte mitogenesis by 1,25-

dihydroxyvitamin D3 (calcitriol). The Journal of Clinical Investigation, 1984. 74(4): p. 1451-1455.


suppress experimental inflammatory bowel disease. European Journal of Immunology, 2005. 35(1): p. 217-224.


with congestive heart failure: a double-blind, randomized, placebo-controlled trial. American Journal of Clinical Nutrition, 2006. 83(4): p. 754-9.


negatively correlated with serum 25(OH)D concentrations in healthy women. Journal of Inflammation, 2008. 5(1): p. 10.


6, tumor necrosis factor-alpha and insulin resistance in overweight and obese subjects. Cytokine, 2012. 60(3): p. 870-4.

156. Ranjit, N., et al., Socioeconomic Position, Race/Ethnicity, and Inflammation in the Multi-

Ethnic Study of Atherosclerosis. Circulation, 2007. 116(21): p. 2383-2390. 157. Dawson-Hughes, B., et al., Estimates of optimal vitamin D status. Osteoporosis

International, 2005. 16(7): p. 713-716. 158. Document : PILKCR-7, 2003-11-25, High sensitivity CRP, and according to Immulate

1000 Operating Manual, Document #600467, rev B1, revision 5.XX August 2003, , D.P.C. (DCP), Editor 2003.


and its relation to BMI among Puerto Ricans Living in the Greater Boston Area. To be submitted, 2013.

206


cytokines production in PBMC modulating intracellular calcium. Transplantation Proceedings, 2001. 33(3): p. 2366-2368.


macrophages from CAPD patients. Kidney International, 2001. 59(1): p. 69–75. 162. Nonn, L., et al., Inhibition of p38 by Vitamin D Reduces Interleukin-6 Production in


Implications for Prostate Cancer Prevention by Vitamin D. Cancer Research, 2006. 66(8): p. 4516-4524.



Examination Survey 2001 to 2006). American Journal of Cardiology, 2012. 109(2): p. 226-30.


Hispanic elders in Massachusetts. J Gerontol B Psychol Sci Soc Sci, 2000. 55(2): p. S108-16.

165. Fantuzzi, G., Adipose tissue, adipokines, and inflammation. The Journal of allergy and clinical immunology, 2005. 115(5): p. 911-9; quiz 920.

166. Van Lente, F., Markers of inflammation as predictors in cardiovascular disease. Clin Chim Acta, 2000. 293(1-2): p. 31-52.

167. Organization, W.H., Noncommunicable diseases country profiles 2011. WHO global

report, 2011. p. 209. 168. Mithal, A., et al., Global vitamin D status and determinants of hypovitaminosis D.

Osteoporosis International, 2009. 20(11): p. 1807-20. 169. Ogden CL, C.M., Kit BK, Flegal KM., Prevalence of obesity in the United States, 2009–

2010. NCHS data brief, 2012, National Center for Health. 170. Flegal Km, C.M.D.K.B.K.O.C.L., Prevalence of obesity and trends in the distribution of

body mass index among us adults, 1999-2010. JAMA, 2012. 307(5): p. 491-497. 171. Renzaho, A.M.N., J.A. Halliday, and C. Nowson, Vitamin D, obesity, and obesity-related

chronic disease among ethnic minorities: A systematic review. Nutrition, 2011. 27(9): p. 868-879.


levels: a population-based study in older men and women. Journal of Clinical Endocrinology & Metabolism, 2005. 90(7): p. 4119-23.


low vitamin D intake in a large US sample of healthy middle-aged men and women. J Steroid Biochem Mol Biol, 2010. 121(1-2): p. 462-6.

174. Norman, A.W., Minireview: vitamin D receptor: new assignments for an already busy

receptor. Endocrinology, 2006. 147(12): p. 5542-8. 175. Tuohimaa, P., Vitamin D and aging. Journal of Steroid Biochemistry & Molecular

Biology, 2009. 114(1-2): p. 78-84. 176. Young, K.A., et al., Association of Plasma Vitamin D Levels with Adiposity in Hispanic

and African Americans. Journal of Clinical Endocrinology & Metabolism, 2009. 94(9): p. 3306-3313.

207


markers of the metabolic syndrome and diabetes in overweight and obesity. Nutrition Journal, 2008. 7: p. 4.


concentrations: considerations and implications. Int J Obes, 2012. 36(3): p. 387- 96. 179. Compher, C. and K.O. Badellino, Obesity and Inflammation: Lessons From Bariatric

Surgery. Journal of Parenteral and Enteral Nutrition, 2008. 32(6): p. 645-647. 180. Li, J., et al., 1α,25-Dihydroxyvitamin D hydroxylase in adipocytes. The Journal of Steroid

Biochemistry and Molecular Biology, 2008. 112(1–3): p. 122-126. 181. Looker, A.C.J., C L. Lacher, D A. Pfeiffer, C M. Schleicher, R L. Sempos, C T.,

Vitamin D status: United States, 2001–2006, in NCHS data brief2011, National Center for Health Statistics: Hyattsville, MD.


in US adults. Nutrition Research, 2011. 31(1): p. 48-54. 183. Flegal, K.M., et al., PRevalence and trends in obesity among us adults, 1999-2008.

JAMA, 2010. 303(3): p. 235-241. 184. Casas, Y.G., et al., Total and regional body composition across age in healthy Hispanic

and white women of similar socioeconomic status. The American Journal of Clinical Nutrition, 2001. 73(1): p. 13-18.

185. Wiklund, P., et al., Abdominal and Gynoid Fat Mass Are Associated with Cardiovascular

Risk Factors in Men and Women. Journal of Clinical Endocrinology & Metabolism, 2008. 93(11): p. 4360-4366.

186. McCarty, M.F., A paradox resolved: the postprandial model of insulin resistance

explains why gynoid adiposity appears to be protective. Medical Hypotheses, 2003. 61(2): p. 173-176.

187. Weishaar, T. and J.M. Vergili, Vitamin D Status Is a Biological Determinant of Health

Disparities. Journal of the Academy of Nutrition and Dietetics, 2013. 113(5): p. 643-651. 188. McKinney, K., C.R. Breitkopf, and A.B. Berenson, Association of race, body fat and

season with vitamin D status among young women: a cross-sectional study. Clinical Endocrinology, 2008. 69(4): p. 535-541.

189. Parikh, S.J., et al., The Relationship between Obesity and Serum 1,25-Dihydroxy Vitamin

D Concentrations in Healthy Adults. Journal of Clinical Endocrinology & Metabolism, 2004. 89(3): p. 1196-1199.

190. Arunabh, S., et al., Body Fat Content and 25-Hydroxyvitamin D Levels in Healthy

Women. Journal of Clinical Endocrinology & Metabolism, 2003. 88(1): p. 157-161. 191. White, J., et al., Precision of single vs bilateral hip bone mineral density scans. Journal of

clinical densitometry : the official journal of the International Society for Clinical Densitometry, 2003. 6(2): p. 159-62.

192. Mazess, R.B., et al., Dual-energy x-ray absorptiometry for total-body and regional bone-

mineral and soft-tissue composition. The American Journal of Clinical Nutrition, 1990. 51(6): p. 1106-12.

193. Jamal-Allial, A.A., J.L. Griffith, and K.L. Tucker, The Longitudinal Association of

Vitamin D Serum Concentrations & Adiposity Phenotypes. To be submitted, 2013. 194. Jacobs, E.J., et al., WAist circumference and all-cause mortality in a large us cohort.

Archives of Internal Medicine, 2010. 170(15): p. 1293-1301.

208

195. Panagiotakos, D.B., et al., The implication of obesity and central fat on markers of

chronic inflammation: The ATTICA study. Atherosclerosis, 2005. 183(2): p. 308-315. 196. Dos Santos, R.E., et al., Relationship of body fat distribution by waist circumference,

dual-energy X-ray absorptiometry and ultrasonography to insulin resistance by

homeostasis model assessment and lipid profile in obese and non-obese postmenopausal

women. Gynecological Endocrinology, 2005. 21(5): p. 295-301. 197. Fontana, L., et al., Visceral Fat Adipokine Secretion Is Associated With Systemic

Inflammation in Obese Humans. Diabetes, 2007. 56(4): p. 1010-1013. 198. Savgan-Gurol, E., et al., Waist to hip ratio and trunk to extremity fat (DXA) are better

surrogates for IMCL and for visceral fat respectively than for subcutaneous fat in

adolescent girls. Nutrition & Metabolism, 2010. 7(1): p. 86. 199. Roger, V.L., et al., Heart disease and stroke statistics--2011 update: a report from the

American Heart Association. Circulation, 2011. 123(4): p. e18-e209. 200. Barcelo, A., et al., Waist circumference, BMI and the prevalence of self-reported diabetes

among the elderly of the United States and six cities of Latin America and the Caribbean. Diabetes Res Clin Pract, 2007. 78(3): p. 418 - 427.

201. McTigue K, L.J.C.V.A. and et al., MOrtality and cardiac and vascular outcomes in

extremely obese women. JAMA, 2006. 296(1): p. 79-86. 202. Bischoff-Ferrari, H.A., et al., Estimation of optimal serum concentrations of 25-

hydroxyvitamin D for multiple health outcomes. The American Journal of Clinical Nutrition, 2006. 84(1): p. 18-28.

203. Zadshir, A., et al., The prevalence of hypovitaminosis D among US adults: data from the

NHANES III. Ethnicity & disease, 2005. 15(4 Suppl 5): p. S5-97-101. 204. Affairs, C.o.S., U.S. Census Bureau 2006. American Community Survey: ACS

Demographic and Housing Estimates

Hispanic Health in the United States. JAMA, 2006. 265(2): p. 248 - 252. 205. Tucker, K., et al., Self-reported prevalence and health correlates of functional limitation

among Massachusetts elderly Puerto Ricans, Dominicans, and non-Hispanic white

neighborhood comparison group. J Gerontol A Biol Sci Med Sci, 2000. 55(2): p. M90 - 97.

206. Lin, H., et al., Hypertension among Hispanic elders of a Caribbean origin in

Massachusetts. Ethn Dis, 2002. 12(4): p. 499-507. 207. Warnecke, R., et al., Approaching health disparities from a population perspective: the

National Institutes of Health Centers for Population Health and Health Disparities. Am J Public Health, 2008. 98(9): p. 1608 - 1615.

208. Organization, W.H., Obesity: preventing and managing the global epidemic., in Report of

a WHO Consultation on Obesity World Health Organization1998, World Health Organization: Geneva, Switzerland.

209. Ross, A.C., The 2011 report on dietary reference intakes for calcium and vitamin D. Public Health Nutr, 2011. 14(5): p. 938-9.

210. Zittermann, A., et al., Vitamin D deficiency and mortality risk in the general population:

a meta-analysis of prospective cohort studies. The American Journal of Clinical Nutrition, 2012. 95(1): p. 91-100.

211. Greene-Finestone, L.S., et al., 25-Hydroxyvitamin D in Canadian adults: biological,

environmental, and behavioral correlates. Osteoporos Int, 2011. 22(5): p. 1389-99.

209

212. Janssen, H.C., et al., Determinants of vitamin D status in healthy men and women aged

40-80 years. Maturitas, 2013. 74(1): p. 79-83. 213. Diaz, V.A., et al., The Association of Vitamin D Deficiency and Insufficiency with

Diabetic Nephropathy: Implications for Health Disparities. The Journal of the American Board of Family Medicine, 2009. 22(5): p. 521-527.

214. Holick, M.F., J.A. MacLaughlin, and S.H. Doppelt, Regulation of Cutaneous Previtamin

D: Photosynthesis in Man: Skin Pigment is not an Essential Regulator. Science, 1981. 211(4482): p. 590-593.

215. Jacques, P.F., et al., Plasma 25-hydroxyvitamin D and its determinants in an elderly

population sample. Am J Clin Nutr, 1997. 66(4): p. 929-36. 216. Zhang, R. and D. Naughton, Vitamin D in health and disease: Current perspectives.

Nutrition Journal, 2010. 9(1): p. 65. 217. Martins, D., et al., Prevalence of Cardiovascular Risk Factors and the Serum Levels of

25-Hydroxyvitamin D in the United States: Data From the Third National Health and

Nutrition Examination Survey. Archives of Internal Medicine, 2007. 167(11): p. 1159-1165.

218. Holick, M.F., et al., Vitamin D2 is as effective as vitamin D3 in maintaining circulating

concentrations of 25-hydroxyvitamin D. J Clin Endocrinol Metab, 2008. 93(3): p. 677-81. 219. LIPS, P., et al., The Effect of Vitamin D Supplementation on Vitamin D Status and

Parathyroid Function in Elderly Subjects. Journal of Clinical Endocrinology & Metabolism, 1988. 67(4): p. 644-650.

220. van Rompay, M.I., et al., Acculturation and Sociocultural Influences on Dietary Intake

and Health Status among Puerto Rican Adults in Massachusetts. Journal of the Academy of Nutrition and Dietetics, 2012. 112(1): p. 64-74.

221. Jorde, R. and G. Grimnes, Vitamin D and metabolic health with special reference to the

effect of vitamin D on serum lipids. Progress in Lipid Research, 2011. 50(4): p. 303-312. 222. Foss, Y.J., Vitamin D deficiency is the cause of common obesity. Med Hypotheses, 2009.

72(3): p. 314-21. 223. Rompay, M., et al., Prevalence of Cardiovascular Disease Risk Factors Among Older

Puerto Rican Adults Living in Massachusetts. Journal of Immigrant and Minority Health, 2011. 13(5): p. 825-833.

210

TABLE 6.1 Descriptive characteristics of men and women in the Boston Puerto Rican Health

Study (BPRHS) 1

Women

(n=1050)

Men

(n=445) P

value

Age (y) 57.3 ± 7.5 56.6 ± 8.0 0.1*

Education (< 8th grade %) 48.0 44.0 0.1†

Poverty (%) 62.0 51.0 0.0003†

BMI (kg/m2) 33.0 ± 7.0 30.0 ± 5.2 <.0001*

Serum Albumin (g/L) 42.5 ± 3.1 43.2 ± 3.6 0.001*

Serum Creatinine (nmol/L) 68.0 ± 27.5 88.0 45.1 <.0001*

Number of Miles Walked/day ‡ 1.0 ± 1.4 1.3 ± 1.6 <.0001*

25(OH)D (nmol/L) 43.1 ± 17.0 43.5 ± 16.5 0.6 *

Vitamin D Supplement use (> 5µg/day)

(%) 30.0 24.0 0.02†

Dietary vitamin D intake•|| (µg/day) 5.3± 2.7 5.2 ± 3.4 0.5 *

Total vitamin D intake•|| (µg/day) 7.3 ± 3.3 6.6 ± 4.2 0.001*

BMI, body mass index; 25(OH)D, 25-hydroxyvitamin D. 1Data are presented as mean ± SD for continuous variables or as % for categorical variables * Values were compared using t-test. † Values were compared using χ2- test ‡ Range = 0-10 miles/day • The recommended dietary allowance (RDA) for vitamin D for men and women aged 19-70 years =15 ug /day [82] ||Adjusted for energy intake using ANOVA (PROC GLM; SAS Institute, Cary, NC).

211

TABLE 6.2 Age-adjusted means of serum 25-hydroxyvitamin D (25(OH)D) concentration stratified by sex for different

characteristics among BPRHS participants, ages 45-75 years

Variables Women Men

n (%)* LSM SEM P value n (%) LSM SEM P

value

Age (y)

45-59 651 (61.5%) 42.0 0.64 281 (63.0%) 44.4 1.00

60-75 407 (38.5%) 45.0 0.81 0.007 164 (37.0%) 42.0 1.33 0.1

Vitamin D Supplement (IU/day) † †‡

≤ 200 738 (70.0%) 41.1 0.6 338 (77.3%) 41.5 0.8

200-400 224 (21.5%) 46.5 1.1 <.0001 90 (20.5%) 51.0 1.6 <.0001

≥ 400 77 (8.5%) 53.1 2.0 <.0001 9 (2.1 %) 39.2 5.36 0.9

<.0001 || <.0001

||

Season of measurement†

Summer (July-September) 356 (34.5%) 44.5 0.9 137 (32.0%) 46.5 1.4

Fall (October- December) 264 (26.0%) 46.0 0.8 0.4 124 (29.0%) 48.7 1.4 0.6

Winter (January-March) 184 (18.0%) 41.1 1.2 0.05 78 (18.0%) 37.5 1.8 <.0001

Spring (April-June) 224 (21.5%) 38.2 1.1 <.0001 91 (21.3%) 37.0 1.6 <.0001

<.0001 || <.0001

||

Body Mass Index (kg/m2) †

≤ 25 114 (11.0%) 44.0 1.5 80 (18.5%) 43.5 2.0

25-30 282 (27.3%) 45.5 0.9 0.8 164 (37.5%) 44.0 1.3 0.9

30-35 307 (29.3%) 44.5 0.9 0.9 130 (30.2%) 43.0 1.5 0.9

≥ 35 307 (29.3%) 44.5 0.9 0.08 130 (30.2%) 43.0 1.5 0.9

212

0.001 || 0.9

||

Smoking Status†

No 832 (79.5%) 44.0 1.1 291 (66.0%) 44.0 1.0

Yes 217 (20.5%) 41.0 0.5 0.01 149 (34.0%) 43.0 1.4 0.6

Alcohol Intake†

No 675 (64.5%) 43.0 0.7 211 (49.0%) 44.5 1.2

Yes 376 (35.5%) 44.0 0.8 0.4 228 (51.0%) 43.0 1.1 0.2

Serum Albumin† (g/L)

< 38 49 (5%) 42.0 2.2 16 (4.0%) 35.0 4.0

≥ 38 964 (95%) 43.1 0.5 0.5 410 (96.0%) 44.0 0.8 0.03

Serum Creatinine (nmol/L) †•

35.5 - 61.0 280 (27.5%) 41.5 1.0 19 (4.5%) 34.5 4.0

62.0 -70.0 298 (29.0%) 42.5 1.0 0.6 46 (11.0%) 40.0 2.4 0.3

71.0 -79.5 223 (22.0%) 45.3 1.1 0.009 92 (21.5%) 42.0 2.0 0.1

80.0 - 698 221 (22.0%) 44.1 1.1 0.1 270 (63.0%) 45.3 1.0 0.01

0.02 || 0.009

||

Miles Walked/day†

0 mile/day 263 (25.0%) 41.3 1.1 52 (12.0%) 40.5 2.4

0- 0.5 mile/day 395 (38.0%) 42.5 0.8 0.5 140 (32.0%) 44.0 1.4 0.3

> 0.5 mile/day 367 (37.0%) 45.0 1.0 0.01 247 (56.3%) 44.0 1.1 0.3

0.01 || 0.4

||

† Age-adjusted means were calculated with LSMEANS ANOVA (PROC GLM; SAS Institute, Cary, NC). Values presented are means || P-values for trend are calculated based on Dunnett adjustment for multiple comparisons (P > 0.05)

213

TABLE 6.3 Correlates1of serum 25(OH)D concentration among Puerto Rican adults stratified by

sex.

Variable

Women (n=935) Men (n=381)

β* SEM P

value† β

* SEM P value†

Age (years) 0.15 10 0.01 -0.03 0.1 0.7

Seasonality

Fall vs. Summer 2.0 1.3 0.15 2.04 2.02 0.3

Winter vs. Summer - 3.0 1.5 0.04 - 9.5 2.3 <.0001

Spring vs. Summer -6.1 1.4 <.0001 - 10.5 2.3 <.0001

BMI (kg/m2) -0.24 0.1 .002 0.04 0.15 0.7

Miles walked/day 0.6 0.3 0.05 0.7 0.5 0.1

Vitamin D supplement (µg/day)

< 5 vs. 10 µg/day -11.0 2.0 <.0001 1.1 5.6 0.8

5-10 vs. 10 µg/day -6.5 2.0 0.002 9.3 5.7 0.1

Dietary vitamin D (µg/day) †

1.0 0.2 <.0001 0.3 0.2 0.2

Serum Albumin (g/L) 0.5 0.2 0.006 0.6 0.2 0.008

Smoking (No vs. Yes) 3.5 1.3 0.008 0.2 2.0 0.8

1β-coefficients, SEM and P-values were calculated using PROC GLM; (SAS Institute, Cary,

NC). All covariates were included in the model simultaneously, including adjustment for serum creatinine, alcohol intake education (< 8th grade %) and poverty. †Dietary vitamin D adjusted for energy intake.

214

TABLE 6.4 The least-square means of serum 25(OH)D concentration for the effect of vitamin

D supplements categories by sex

Women (n=935) Men (n=381) β SEM β SEM

≤ 200 IU/day 40.25a 0.70 40a 1.0

200-400 45.0b 1.14 48.3ab 1.81

≥ 400 IU/day 51.0c 1.85 37ac 5.40

β coefficients, SEM and P-values were calculated using PROC GLM; (SAS Institute, Cary, NC). The model was adjusted for age, BMI, season, dietary vitamin D, miles walking/day, serum albumin, serum creatinine, smoking, alcohol intake education (< 8th grade %) and poverty.

215

TABLE 6.5 The least-square means of serum 25(OH)D concentration and seasons categories by

age groups

40-59 Years 60-75 Years β

* SEM β* SEM

Summer 50 1.50 46 1.5 Fall 50.3 1.15 48.5 1.32

Winter 41.3 1.35 45.2 1.75 Spring 38.4 1.32 43.4 1.65

β coefficients, SEM and P-values were calculated using PROC GLM; (SAS Institute, Cary, NC). The model was adjusted for sex, BMI, dietary vitamin D, miles walking/day, serum albumin, serum creatinine, smoking, alcohol intake education (< 8th grade %) and poverty.

216

TABLE 6.6 Major Dietary Sources of Vitamin D intake among Puerto Rican adults *

Rank Food Groups % contribution of dietary vitamin D intake

1 Milk (whole, 2%,1% & milk products) 32.4 2 Supplements 28.7 3 Fish 16.2 4 Eggs 4.5 5 Cereal, Ready-to-eat, Cold 4.3 6 Beef 2.3 7 Processed Meat, Sausage, Frank 2.1 8 Chicken/Turkey 2.0 9 Pork 1.3 10 Ice cream, Sherbet, Frozen Yogurt 1.2 11 Yogurt 1.1 12 Cheese 0.9

* Food groups listed were the ones that contributed more than 0.8% of dietary vitamin D intakes.

217

FIGURE 6.1 Prevalence of low serum 25-hydroxyvitamin D concentration among Puerto Rican

men (n=426) and women (n=1015). Serum 25(OH)D concentration was categorized: deficient ≤

50 nmol/l, insufficient 50-75 nmol/L, and sufficient ≥ 75 nmol/L. Values are percentage of

individuals.

0

10

20

30

40

50

60

70

80

Men Women

Per

cen

tage

%

<50

50 - 75

75 nmol/L

218

FIGURE 5.2 Association between the median intake of total vitamin D as defined by the RDA

and serum 25(OH)D concentration (nmol/L) for BPRHS participants (n=1433). The vitamin D

RDA is 15 µg/day = 600 IU/day. Data points are least square means for serum 25(OH)D,

adjusted for age, sex, and season, plotted against category of the RDA of vitamin D intake (P-

trend <0.0001). RDA categories are significantly different from referent group ≤ ½ RDA

(Dunnett adjustment for multiple comparisons; P < 0.05).

30

35

40

45

50

55

60

≤ ½ RDA ½ RDA - RDA ≥ RDA

Ser

um

25(O

H)D

nm

ol/

L

Median intake of vitamin D by µg/day the RDA

219

* *

FIGURE 6.3 Association between vitamin D intake from food only sources (µg/day) and serum

25(OH)D concentration (nmol/L) among non-supplement users of the BPRHS participants

(n=1036). Data points are least square means, adjusted for age, sex, and season, plotted at the

median for each quintile category of energy-adjusted vitamin D intake from food (P-trend =

0.0007). Dietary intake of vitamin D equal to 5.28 µg/day and higher contributed to a

significantly higher serum 25(OH)D concentration (nmol/L) compared to the referent group

5.23µg/day (Dunnett adjustment for multiple comparisons; P < 0.05).

30

35

40

45

50

55

5.23 5.26 5.28 5.31 5.37

Ser

um

25(O

H)D

con

cen

trati

on

(n

mol/

L)

Dietary intakes of vitamin D µg/day by quintiles

*

220

FIGURE 6.4 Mean serum 25(OH)D concentration by BMI categories of Puerto Rican women

(n=990), adjusted for age, season, dietary vitamin D and vitamin D supplement use (P-trend =

0.001). BMI ≥40 is significantly different from the referent group BMI ≥ 25; (Dunnett

adjustment for multiple comparisons; P < 0.05).

35

37

39

41

43

45

47

49

51

53

≤ 25 25-30 30-35 35-40 ≥ 40

Ser

um

25(O

H)D

con

cen

trati

on

(n

mol/

L)

BMI (kg/m2)

*

Documents

Serum 25(OH)D concentration & cardiovascular disease risk …rx... · 2019-02-13 · SERUM 25(OH)D CONCENTRATION & CARDIOVASCULAR DISEASE RISK ASSCOCIATIONS AMONG OLDER PUERTO RICANS